Embed Size (px)
Citation preview
Biological and Integrated Control of Water Hyacinth,
Eichhornia crassipes
Proceedings of the Second Meeting of the Global Working Group for the Biological and Integrated Control of Water
Hyacinth, Beijing, China, 9–12 October 2000
Editors
: M.H. Julien, M.P. Hill, T.D. Centerand Ding Jianqing
Sponsored by
: International Organization for Biological Control
Organised by
: Institute of Biological Control, Chinese Academy of Agricultural Sciences
Supported by
: National Natural Scientific Foundation of ChinaChinese Academy of Agricultural Sciences
Australian Centre for International Agricultural ResearchCanberra 2001
The Australian Centre for International Agricultural Research (ACIAR) was estab-lished in June 1982 by an Act of the Australian Parliament. Its primary mandate isto help identify agricultural problems in developing countries and to commissioncollaborative research between Australian and developing country researchers infields where Australia has a special research competence.
Where trade names are used this constitutes neither endorsement of nor discrim-ination against any product by the Centre.
© Australian Centre for International Agricultural Research, GPO Box 1571,Canberra ACT 2601.
Julien, M.H., Hill, M.P., Center, T.D. and Ding Jianqing, ed. 2001. Biological andIntegrated Control of Water Hyacinth,
Eichhornia crassipes.
Proceedings of theSecond Meeting of the Global Working Group for the Biological and IntegratedControl of Water Hyacinth, Beijing, China, 9–12 October 2000. ACIARProceedings No. 102, 152p
ISBN 1 86320 319 2 (print)1 86320 320 6 (electronic)
ACIAR PROCEEDINGS
This series of publications includes the full proceedings of researchworkshops or symposia organised or supported by ACIAR. Numbers inthis series are distributed internationally to selected individuals andscientific institutions.
3
Contents
Address of Welcome 5
Editorial 6
Biological Control of Water Hyacinth with Arthropods: a Review to 2000 8
M.H. Julien
Biological Control of Water Hyacinth by Using Pathogens: Opportunities, Challenges, and Recent Developments 21
R. Charudattan
Water Hyacinth in China: Its Distribution, Problems and Control Status 29
Ding Jianqing
,
Wang Ren
,
Fu Weidong and Zhang Guoliang
Biological Control Initiatives against Water Hyacinth in South Africa: Constraining Factors, Success and New Courses of Action 33
M.P. Hill
and T. Olckers
Recent Efforts in Biological Control of Water Hyacinth in the Kagera River Headwaters of Rwanda 39
T.M. Moorhouse
,
P. Agaba and T.J. McNabb
Ongoing Activities in the Biological Control of Water Hyacinth in Egypt 43
Y.H. Fayad
,
A.A. Ibrahim
,
A.A. El-Zoghby
and F.F. Shalaby
Progress with Biological Control of Water Hyacinth in Malawi 47
P.M. Phiri
,
R.K. Day
,
S. Chimatiro, M.P. Hill, M.J.W. Cock
,
M.G. Hill and E. Nyando
Biological Control of Water Hyacinth by a Mycoherbicide in Egypt 53
Y.M. Shabana, M.A. Elwakil and R. Charudattan
IMPECCA: an International, Collaborative Program to Investigate the Development of a Mycoherbicide for Use against Water Hyacinth in Africa 57
R. Bateman
Fungi Associated with
Eichhornia crassipes
(Water Hyacinth) in the Upper Amazon Basin and Prospects for Their Use in Biological Control 62
H.C. Evans and R.H. Reeder
A Water Hyacinth Resource Manual 71
G. Hill
and R. Day
Water Hyacinth Information Partnership for Africa and the Middle East 72
L.A. Navarro
Can Competition Experiments Be Used to Evaluate the Potential Efficacy of New Water Hyacinth Biological Control Agents? 77
T.D. Center
,
T.K. Van
,
and M.P. Hill
4
How Safe Is the Grasshopper
Cornops aquaticum
for Release on Water Hyacinth in South Africa? 82
I.G. Oberholzer and M.P. Hill
Establishment, Spread and Impact of
Neochetina
spp. weevils (Coleoptera: Curculionidae) on Water Hyacinth in Lake Victoria, Kenya 89
G.S. Ochiel
,
S.W. Njoka
,
A.M. Mailu and W. Gitonga
Water Hyacinth Population Dynamics 96
J.R. Wilson
,
M. Rees
,
N. Holst
,
M.B. Thomas and G. Hill
Current Strategies for the Management of Water Hyacinth,
Eichhornia crassipes
on the Manyame River System in Zimbabwe 105
G.P. Chikwenhere
Biomass and Productivity of Water Hyacinth and Their Application in Control Programs 109
E.L. Gutiérrez
,
E.F. Ruiz
,
E.G. Uribe
and J.M. Martínez
Water Hyacinth Control through Integrated Pest Management Strategies in Tanzania 120
G. Mallya
,
P. Mjema and J. Ndunguru
Integrated Control of Water Hyacinth (
Eichhornia crassipes
) on the Nseleni/Mposa Rivers and Lake Nsezi, Kwa Zulu-Natal, South Africa 123
R.W. Jones
Preliminary Assessment of the Social, Economic and Environmental Impacts of Water Hyacinth in Lake Victoria Basin and Status of Control 130
A.M. Mailu
Biological Control of Water Hyacinth by
Neochetina eichhorniae
and
N. bruchi
in Wenzhou, China 140
Lu Xujian
,
Fang Yongjun
,
Song Darong
and Xia Wanqing
Session Summaries 141
Sessions 1 and 2. Keynote papers and Biological control—general 142Session 3. Biological control—pathogens 146Session 4. General 148Session 5. Biological control—insects 149Session 6. Integrated management 151
5
Address of Welcome
Good morning, ladies and gentlemen
On the occasion of this splendid autumn season in Beijing, it is a great pleasurefor all of us to greet the opening of the Second Global Working Group Meeting forthe Biological and Integrated Control of Water Hyacinth, held under the auspicesof the International Organization of Biological Control (IOBC). On behalf of theBiological Control Institute of the Chinese Academy of Agricultural Sciences,please allow me to extend our warm congratulations. I sincerely wish the meetinga great success.
As is known to all, water hyacinth is one of the most dangerous weeds in theworld, causing great damage to agriculture, aquatic production, tourism and theenvironment in over 40 countries including China. Currently in China, water hya-cinth grows in 17 provinces, and millions of dollars are spent on its control everyyear. Although China has made great efforts and achieved remarkable progress inthe biological and integrated control of water hyacinth, the weed is still spreadinginto new regions at an alarming speed. As this working group meeting provides agood opportunity for mutual exchange of information and experiences in the fieldof water hyacinth control among the delegates and scientists from various coun-tries, I am sure that a successful meeting will not only promote the research workon water hyacinth control in China, but also help advance research activities onglobal water hyacinth control into a new phase. Therefore, it is essential that theworking group meeting be held regularly so that scientists and experts from dif-ferent countries and regions can get together and cooperate in finding good solu-tions to the worldwide problem of water hyacinth.
Finally, I’d like to wish everyone a nice stay in China.
Thank you.
Professor Yang Huaiwen,Director of the Biological Control Unit,
Chinese Academy of Agricultural Sciences
6
Editorial
These are the papers presented at the Second Global Working Group Meeting forthe Biological and Integrated Control of Water Hyacinth, held in Beijing, China inOctober 2000 under the auspices of the International Organization for Biologicaland Integrated Control of Noxious Animals and Plants (IOBC). The meetingbrought together 31 delegates from 11 countries with the common purpose ofidentifying suitable biological and integrated control options for water hyacinth.
These proceedings represent the current status of work on water hyacinth world-wide and include new research initiatives, overviews of water hyacinth implemen-tation programs in various countries and a proposal for a mechanism to facilitatethe dissemination of information on water hyacinth through a clearinghouse. Thepapers, which were each refereed by at least two of the editors, were presentedunder a series of themes. The salient points from each of the papers were then sum-marised at the end of the theme session. These summaries are included in the pro-ceedings. Nevertheless, much of the discussion that occurred at these workshopsis not recorded here. The address list appended to the proceedings will, it is hoped,stimulate further interaction between the delegates. We are very grateful to theAustralian Centre for International Agricultural Research (ACIAR) for sup-porting the preparation and publication of these proceedings.
One of the roles of this working group is to identify further research needs onwater hyacinth. From the presentations and discussions the following ideasemerged as requiring further investigation.
• The impact of cold climates on the success of biological control. Investigationof the thermal tolerance of the natural enemies used was suggested, and of thevalue of collecting biological control agents from climatically similar localities.Also, studies of the impact of releasing large numbers of healthy, fertile femalesthrough the winter to obviate the lag time in population build-up of the weevilsfollowing cool winters were suggested.
• The use of plant competition studies between water hyacinth and other aquaticplants as an indicator of how effective particular agents are.
• The compatibility of the different control options that could be used inintegrated management.
• The compatibility of each biological control agent with each of the herbicideslikely to be used and their surfactants.
• The selection of suitable locations and undertaking of integrated managementof water hyacinth where biological control is the base technique. This has beendone in South Africa in a temperate climate (Jones, these proceedings). TheKafue River, where various agents are established but control has not beensuccessful, offers an opportunity in the tropics.
• Identification and conduct of surveys for additional natural enemies (bothinsects and pathogens) in new areas in the region of origin of water hyacinth.
• The interactions between the insect natural enemies with the pathogen naturalenemies.
7
• The development of mycoherbicide for water hyacinth. It is hoped that theIMPECCA project (Bateman, these proceedings) will achieve this goal.
• Quantification of the contribution of
Orthogalumna
to biological control in thefield. Studies in Malawi may be the first step in this.
• Quantification of the impact of
Eccritotarsus
in the field.
The workshop closed with a general meeting of the working group (the partici-pants). During the meeting it was suggested that a mission statement for theworking group be developed and this was done (see below). It was also decidedthat the next meeting should be held in Uganda on the shores of Lake Victoria inearly August 2002.
Mission Statement
The mission of the IOBC Working Group for the Biological andIntegrated Control of Water Hyacinth is to promote better managementof water hyacinth through:• facilitation of interactions,• dissemination of information, and• identification of research needs.This will be achieved by:• holding a meeting every 2 to 3 years,• publishing the meeting proceedings, a water hyacinth newsletter and
maintaining web site, and• supporting activities that contribute to better management of water
hyacinth.
Meeting participants
Front row (left to right): Wu Zhenquan, Chen Ruoxia, Joseph Ndunguru, Gasper Mallya,Raghavan Charudattan, Fu Weidong, Ding Jianqing, Peter Mjema and XiaShanlong.
Back row (left to right): Lu Qingguang, Wang Qinghai, Sun Junmao, Ma Ruiyan, TomMoorehouse, Tom McNabb, Eric Gutiérrez, Garry Hill, Yahia Fayad, Ted Center,Roy Bateman, Lius Navarro, Harry Evans, Richard Shaw, Godfrey Chikwenere,John Wilson, Roy Jones, Mic Julien, Andrew Mailu and Martin Hill.
8
Biological Control of Water Hyacinth with Arthropods: a Review to 2000
M.H. Julien
*
Abstract
Water hyacinth, native to the Amazon River, invaded the tropical world over the last century and has becomean extremely serious weed. The search for biological control agents began in the early 1960s and continuestoday. Six arthropod species have been released around the world. They are: two weevils,
Neochetina bruchi
and
N. eichhorniae
; two moths,
Niphograpta albiguttalis
and
Xubida infusellus
; a mite
Orthogalumnaterebrantis
; and a bug
Eccritotarsus catarinensis
. The mite and
X. infusellus
have not contributed to control andthe bug is under evaluation following recent releases in Africa. The two weevils and the moth
N. albiguttalis
have been released in numerous infestations since the 1970s and have contributed to successful control of theweed in many locations. It is timely to assess their impact on water hyacinth and, to help in planning futurestrategies, to identify the factors that contribute to or mitigate against successful biological control. Althoughthe search for new agents continues, and as a result biological control will likely be improved, this techniquealone is unlikely to be successful in all of the weed’s habitats. It is important that whole-of-catchmentmanagement strategies be developed that integrate biological control with other control techniques. The aimsof such strategies should be to achieve the best possible control using methods that are affordable andsustainable; hence the need to develop strategies using biological control as the base component.
W
ATER
hyacinth apparently became a problem in theUSA following its distribution to participants in the1884 New Orleans Cotton Exposition. By the early1900s it was widespread in the southern states. Duringthe same period it spread through the tropics of othercontinents and now reaches around the world and northand south as far as the 40° latitudes (Center 1994).More recently it spread into the many waterways ofAfrica and has expanded rapidly, probably in responseto high nutrient conditions, to cause serious problems.To combat the problems caused by the weed, efforts tocontrol its spread and to reduce its biomass have beenmany and varied and include weed managementmethods such as physical removal, application of her-bicides and release of biological control agents. Utili-
sation of the weed for commercial and subsistencepurposes has also been widely considered. It is nowgenerally recognised that physical and chemical con-trols have very limited application in most countriesbecause of their high cost and low sustainability. Utili-sation has never developed into sustainable activitiesother than localised cottage industries or to supportvery poor communities in subsistence existences suchas the production of biogas. Only small amounts ofwater hyacinth can be utilised in such activities, whichshould never be confused with control. Neither shouldthe potential for utilisation prevent the implementationof control strategies (Harley et al. 1996; Julien et al.1996). The cost of water hyacinth to communities faroutweighs the benefits that might occur through utili-sation. In general, even when the weed is successfullymanaged there is likely to be sufficient present tosupport the small-scale utilisation activities that persist.
* CSIRO Entomology, 120 Meiers Road, Indooroopilly,Queensland 4068, Australia. Email: [email protected]
9
The one control technique that continues to showpromise, can be developed further, is affordable, envi-ronmentally friendly and above all sustainable, is bio-logical control. The remainder of this paper is a reviewof the activities and results of biological control ofwater hyacinth using arthropods, and includes a dis-cussion of the attributes and limitations of this tech-nique. A review of biological control of water hyacinthusing pathogens is presented separately in this volume(Charudattan 2001).
Exploration for Natural Enemies
Surveys for natural enemies of water hyacinth for useas biological control agents began in 1962 and havecontinued until recently. A brief chronology, summa-rised largely from Center (1994), follows.
• Mr A. Silveira-Guido conducted the first surveys inUruguay in 1962 to 1965. He found the moth
Xubida
(
Acigona
)
infusellus
, two weevil species
Neochetina eichhorniae
and
Neochetina bruchi
, themite
Orthogalumna terebrantis
and the grasshopperCornops aquaticum, among other species.
• Biology and host range studies were conducted on anumber of these agents at the USDA-ARSlaboratory at Buenos Aires. This laboratory was setup in 1962 to work on alligator weed and from 1968studies focused largely on water hyacinth.
• During 1968 surveys were conducted by F. Bennettand H. Zwölfer of CIBC, now CABI Biosciences, inGuyana, Surinam and Brazil. To the list of speciesthey added the petiole-tunnelling moth Niphograpta(Sameodes) albiguttalis, the petiole-boring fliesThrypticus spp., and an unnamed mirid bug.
• D. Mitchell and P. Thomas conducted surveys inUruguay, Brazil, Guyana and Trinidad but did notextend the list of known phytophages.
• Bennett surveyed the West Indies, Belize andFlorida USA in the late 1960s and foundO. terebrantis and the stem-boring moth Belluradensa.
• Surveys were also carried out in India in the early1960s by Rao, and in Indonesia in the mid 1970s byMangoendihardjo and Soerjani.
• In 1969 R. Gordon and J. Coulson conductedsurveys in Florida, Louisiana and Texas, USA, andfound O. terebrantis and B. densa.
• In 1981 Bennett surveyed Mexico, findingX. infusellus, N. eichhorniae, C. aquaticum andO. terebrantis.
• In 1989 Stephan Neser, PPRI, collected the miridEccritotarsus catarinensis in Santa Catarina State,Brazil (Hill et al. 1999). This may have been the bugrecorded by Bennett and Zwolfer during their 1968surveys in Guyana, Surinam and Brazil.
• In 1999 a survey was conducted by M. Hill, PPRISouth Africa, H. Cordo and T. Center, USDA-ARS,and H. Evans and D. Djeddour, CABI Biosciences,into the upper reaches of the Amazon River in Peru(M. Hill, pers. comm. 1999).
The native range of water hyacinth is widelyreferred to as South America. Using the variations inflower morphology, Barrett and Forno (1982) sug-gested that it was more accurately the Amazon Basin.Recent surveys suggest that the centre of origin forwater hyacinth may be the upper reaches of theAmazon River and its tributaries. The reasons are thatthe widest diversity of fauna associated with the planthas been found in that area, and the floating habit ofthe plant probably evolved to withstand rapid fluctua-tions in water level that occur in the upper AmazonRiver (T. Center and M. Hill, pers. comm. 1999).
The range of surveys provided lists of fauna relatedto the weed. From these lists arthropods and pathogenshave been selected for further studies. The selectionprocess relies initially on the observations and judg-ment of the surveying scientists. The host ranges ofthose selected are observed in the field and studied inthe laboratory. Those showing a narrow host range arethen subjected to host-specificity tests to determine thesafety of releasing them in the exotic range of theweed. The listing and selection of potential agents is acontinuous process that occurs while surveys continueand as new information becomes available about thefauna. The most recent list was presented to the lastInternational Organisation for Biological ControlWater Hyacinth Workshop in 1998 (Cordo 1999),where three levels of priority were assigned to groupsof potential agents. The first priority group listed thefour agents that have been released for some time.They are: the weevils N. eichhorniae and N. bruchi;the moth N. albiguttalis; and the mite O. terebrantis.The second priority group included agents that haverecently been released—E. catarinensis andX. infusellus—and others recently or currently understudy including C. aquaticum, B. densa, the moth Par-acles (Palustra) tenuis and the flies Thrypticus species.The third priority list included a list of nine organisms(eight insects and a mite) about which little is known.The second and third priority lists may change as aresult of the recent and proposed surveys in Peru.
10
Biological Control Agents
Six arthropod biological control agents have beenreleased around the world (Table 1). Five are insects(two weevils, two moths and a sucking bug), and oneis a mite. The two weevils, N. bruchi and N, eichhor-niae, and one of the moths, N. albiguttalis, have beenreleased widely since 1971 in 30, 32 and 13 countries,respectively, while the others, the mite O. terebrantis,the moth X. infusellus and the bug E. catarinensis,have been released in fewer countries: 2, 3 and 6,respectively. The mite was first released in 1971 whilethe other two were first released in 1996.
Neochetina bruchiBiology
Small whitish eggs are laid into the petioles, ofteninto insect chew holes, singly or several together. Eggshatch in about seven days and don’t hatch below 15oC.Larvae tunnel inside the petioles towards the base andinto the crown where they often feed on developingaxillary buds. A number of larvae may feed in thesame petiole or crown and they may move betweenpetioles. Larvae have three instars and developmenttakes about 33 days, the rate of development beingtemperature and nutrient dependent. Final instar larvaeexit the crown and move to the roots and construct acircular cocoon using excised root hairs, attached to alarger root. Pupal development takes about 20 days.Adult beetles are 4–5 mm long and tan brown incolour. They are nocturnal and remain concealed in thecrown of the plant. They feed externally on the epi-dermal tissues of the leaves, forming characteristicfeeding scars. Adults also feed preferentially on thenarrow upper part of the petiole of the first and secondleaves. Most eggs are laid within five weeks of emer-gence, and between about 300 and 700 eggs have beenrecorded per female. Adults may live to nearly 100days. The generation time has been recorded atbetween 72 and 96 days. The optimum temperature forfeeding and oviposition is about 30°C. High tempera-ture and low humidity may decrease egg productionand reduce adult survival, while low temperature,probably below about 15°C, arrests development, pre-vents population increase and decreases survival(Cordo and DeLoach 1976; Julien et al. 1999)
DamageThe damage caused by this insect and by N. eichhor-
niae (see below) is similar. Adult feeding scars, whennumerous, debilitate the plant by removing extensiveproportions of epidermal tissue thus increasing water
loss and exposing the plant to attack by pathogens.Extensive feeding around the upper petiole may girdlethe petiole and kill the lamina above. Larval tunnellingin the lower petiole and crown damages tissues andbuds, initially preventing flowering. As damageincreases, plant growth rate is reduced and the produc-tion of new leaves and new stolons is reduced. Plantsize (height, weight, size of leaves, size of stolons)declines. Internal damage to plant tissues results inrotting of the lower petioles, waterlogging of thecrown and gradual sinking of the plant so that thecrown is several centimetres below the surface of thewater. In time the plant dies, most sinking, thoughsome may remain as a floating mass. The process fromrelease of the weevils to plant death takes years, theduration depending on a combination of factors, suchas temperature, nutrient status of the weed, climate,hydrology of the catchment, and number of healthyinsects released.
Releases The first recorded release was in 1974, in the USA.
Neochetina bruchi was recorded in Mozambique in1972 but there is no record of how it got there. It hasbeen released in 30 countries, is not known to be estab-lished in four and recent releases in three others areunder evaluation (Table 2). This weevil is contributingto control of the weed in 11 countries where the initialreleases were made between 1974 and 1996. It isestablished and under evaluation in four other coun-tries and, unfortunately, there are no post-releaseassessments for seven countries. Neochetina bruchiwas distributed within Argentina in 1974 and inBolivia (year unknown) to areas where the weed hadbecome a problem (Julien and Griffiths 1998; Julien etal. 1999). It was released in The Republic of Congo in1999 (IITA 2000), and in Egypt (Fayad et al. 2001)and Rwanda during 2000 (Moorhouse et al. 2001).
Neochetina eichhorniaeBiology
This insect’s small whitish eggs are more slenderand softer than those of N. bruchi. They are laid singlybeneath the epidermis of the leaves, petioles andligules. Eggs hatch in about 10 days and will not hatchat temperatures below 20°C. Larvae have three instarsand tunnel inside petioles towards and into the crown.A number of larvae may exist in the same petiole orcrown where they damage axillary buds. The rate ofdevelopment of larvae is dependent on temperatureand nutrition and takes 60–90 days. Construction of acocoon, about 5 mm diameter, and pupal development
11
Table 1. Countries (total 34) where biological control agents have been released on water hyacinth and the dates ofinitial releases. Data modified from Julien and Griffiths (1998)
Neochetina bruchi
Neochetina eichhorniae
Niphograpta albigutallis
Eccritotarsus catarinensis
Orthogalumna terebrantis
Xubida infusellus
Australia 1990 1975 1977 1981; 1996f
Benin 1992 1991 1993 1999h
China 1996 1996 2000a
a. Ding et al. (2001).
Congo 1999h 1999h
Cuba 1995
Egypt 2000b 2000b
Fiji 1977
Ghana 1994 1994 1996
Honduras 1989 1990
India 1984 1983 1986
Indonesia 1996 1979
Kenya 1995 1993
Malawi 1995 1995 1996 1996
Malaysia 1992 1983 1996
Mexico 1995 1972
Mozambique 1972 1972
Myanmar 1980
Nigeria 1995 1993
Panama 1977 1977
Philippines 1992 1992
PNG 1993 1986 1994 1996
Rwanda 2000d 2000d
Solomon Islands 1988
South Africa 1989 1974 1990 1996
Sri Lanka 1988
Sudan 1979 1978 1980
Taiwan 1993 1992
Tanzania 1995 1995
Thailand 1991 1979 1995 1999
Uganda 1993 1993
USA 1974 1972 1977
Vietnam 1996 1984
Zambia 1997c 1971; 1996 1971; 1997g 1997c 1971
Zimbabwe 1996 1971 1994 1999e
Totals 30 32 13 6 2 3
b. Fayad et al. (2001).c. M. Hill (pers. comm., 2000).d. Moorhouse et al. (2001).e. G. Chikwenhere (pers. comm., 2000).f. Failed to persist after releases in 1981 and was imported and released again in 1996 (Julien and Stanley 1999).g. Initial releases did not establish and it was released again in 1997 (M. Hill, pers. comm. 2000).h. IITA (2000).
12
Table 2. Neochetina bruchi: status of releases for each country. Data modified from Julien and Griffiths (1998)
Established Control
Year released
No Unknown Under evaluation
Yes No Yes Under evaluation
Unknown
Panama 1977
Philippines 1992
Taiwan 1993
Zambiaa
a. M. Hill (pers. comm., 2000).b. Fayad et al. (2001).c. Moorhouse et al. (2001).d. IITA (2000).
1997
Congod 1999
Egyptb 2000
Rwandac 2000
Malaysia 1992
Benind 1992
Australia 1990
India 1984
Kenya 1995
PNG 1993
Sudan 1979
Tanzania 1995
Thailand 1991
Uganda 1993
USA 1974
Zimbabwe 1996
China 1996
Malawi 1995
Mexico 1995
South Africa 1989
Cuba 1995
Ghana 1994
Honduras 1989
Mozambique 1972
Nigeria 1995
Indonesia 1996
Vietnam 1996
13
are similar to that for N. bruchi. Adults are 4–5 mmlong, slightly smaller than N. bruchi, and are colouredmostly grey. They feed nocturnally and hide in thecrown during daylight. Adults feed externally on theepidermal tissue of the leaves and upper petioles pro-ducing feeding scars indistinguishable from thosecaused by N. bruchi feeding. The generation time islonger than for N. bruchi, 96–120 days. Adult lon-gevity has been recorded at 140 and 300 days and eggsper female at 5–7 per day and 891 total (Cordo andDeLoach 1976; Julien et al. 1999). Neochetina eich-horniae is less dependent on good quality plants fordevelopment than N. bruchi. Consequently, the rela-tive abundance varies between sites; more N. eichhor-niae at sites with lower quality water hyacinth, andvice versa.
The two Neochetina species can be readily distin-guished in the adult stage. In N. bruchi two dark markson the elytra are equal in length, are relatively shortand are located midway along the elytra. The elytrafurrows are broader and have comparatively shallowcurvature. New adults have scale coloration that formsa ‘v’ on the elytra. This mark fades with age. In com-parison, the two elytra marks on N. eichhorniae arelonger, not equal in length and tend to occur closer tothe front of the elytra. The elytra furrows are narrowwith strong curvature. There is no ‘v’ pattern on theelytra (Julien et al. 1999).
Damage
This weevil damages water hyacinth in a similarway to N. bruchi (see above). An important differenceis that N. bruchi populations develop better undereutrophic conditions (Heard and Winterton 2000) and,in polluted waterways, may complement the damageby N. eichhorniae.
Releases
The first releases were in 1971 in Zambia and Zim-babwe. Thereafter it was released in another 32 coun-tries (Table 3). This insect is established in all but sixcountries and three of these were recent releases and areunder evaluation. It contributes to control the weed in13 countries where releases were made between 1971and 1995. It is being evaluated in two others and thereis no post-release information about control from sevencountries. Neochetina eichhorniae was distributed inBolivia (year unknown) to areas where the weed hadbecome a problem (Julien and Griffiths 1998; Julien etal. 1999). It has been released in The Republic of Congoin 1999 (IITA 2000), and in Egypt (Fayad et al. 2001)and Rwanda during 2000 (Moorhouse et al. 2001).
Niphograpta albiguttalisBiology
Eggs are creamy white, 0.3 mm diameter and are laidsingly or in small groups in leaf tissue, particularly atinjury and feeding sites. Hatching occurs in 3–4 days.The five larval instars develop over 16–21 days. Larvaefeed externally initially and after one or two days theytunnel into the petiole and feed below the epidermiscausing characteristic ‘windows’. As larvae grow theytunnel deeper into the petiole tissues and into the centralrosette of the plant. They may move between petiolesand several larvae may feed in the same petiole. Larvaeare rarely found in older, tougher plants or petioles, butprefer younger, tender material, characteristic of thesmall bulbous plants that grow on the edge of water hya-cinth infestations. Pupation occurs in a chamber chewedin a relatively undamaged portion of petiole with atunnel leading to the leaf epidermis where a thin windowis left for protection across the emergence exit. Pupationoccurs within a white cocoon and takes about 5–7 days.The adult moves up the emergence tunnel and exitsthrough the ‘window’ in the epidermis. Adults are 6–10mm long with a wingspan of 17–25 mm. Colour is var-iable from golden yellow to charcoal grey, with brown,black and white markings. Mating occurs soon afteremergence and oviposition begins soon afterwards.Some 70% of eggs are laid during the second and thirdnights and moths live for 4–9 days. Females lay 370eggs on average. The life cycle takes 21–28 days. Forgreater detail see Bennett and Zwolfer (1968), DeLoachand Cordo (1978), Center (1981) and Harley (1990).
Damage Early larval tunnelling causes necrosis and water-
logging of internal tissues. Small, dark spots occur onthe surface of the petiole. Larger larvae cause severe,internal damage causing petioles and leaves to wilt,turn brown and rot. When damage destroys the apicalbud, growth is prevented and ramet death occurs.However, axillary buds may continue to develop and,unless attacked by the moth, will replace the deadramet. The adult moths disperse rapidly, up to 4 km perday. Severe local damage to water hyacinth may occur,but overall the damage is patchy as adults tend to ovi-posit on healthy young, tender plants. Quantifying theimpact of this moth on weed populations is extremelydifficult. Its role in biological control is thought to bein slowing the rate of expansion of mats by reducingnew growth along the expanding edges. It could alsoplay an important role in reducing the rate of invasionby preferentially attacking rapidly growing plants thatare typical of invasion and regrowth areas.
14
Table 3. Neochetina eichhorniae: status of releases for each country. Data modified from Julien and Griffiths (1998)
Established Control
Year released
No Unknown Under evaluation
Yes No Yes Under evaluation
Unknown
Philippines 1992
Taiwan 1992
Vietnam 1984
Congo3 1999
Egypta
a. Fayad et al. (2001).b. Moorhouse et al. (2001).c. IITA (2000).
2000
Rwanda2 2000
Fiji 1977
Indonesia 1979
Mexico 1970
Sri Lanka 1988
Australia 1975
Benin 1991
India 1983
Kenya 1993
Nigeria 1993
PNG 1986
South Africa 1974
Sudan 1978
Tanzania 1995
Thailand 1979
Uganda 1993
USA 1972
Zimbabwe 1971
China 1996
Malawi 1995
Ghana 1994
Honduras 1990
Malaysia 1983
Mozambique 1972
Myanmar 1980
Solomon Islands 1988
Zambia 19711996
15
Releases
This moth was first released in Zambia in 1971 andhas been released in a total of 13 countries (Table 4). Itis established in six countries, contributes to control intwo and is being evaluated in three others. Althoughnot deliberately released there, this insect has beenrecorded in Cuba.
Eccritotarsus catarinensis
Biology
Eggs are inserted into the leaves just below the sur-face. Four nymphal instars are gregarious and feed onthe surface of the laminae with the adults. Nymphs arepale, while adults, which are 2–3 mm long, have darkbodies and pale wings with dark markings. The devel-opment of the immature stages (egg to adult) takes 22days and adults live for about 50 days (Stanley andJulien 1999; Hill et al. 1999).
Damage
Feeding by the nymphs and adults of this small,sucking bug causes chlorosis of the laminae. Withsevere damage, photosynthesis and therefore growthand reproduction of the weed could be reduced.
Releases
E. catarinensis was recently studied in South Africaand Australia. It was released in South Africa in 1996(Hill et al. 1999). It has also been released in Malawi in1996 (Julien and Griffiths 1998), Zambia in January1997 (Hill 1997), Zimbabwe in May 1999 (G. Chik-wenhere, pers. comm. 2000), Benin in June 1999 (O.Ajuonu, pers. comm. 2000), and in China during thespring of 2000 (Ding et al. 2001). It was not released inAustralia because of its potential to damage nativeMonochoria species (Stanley and Julien 1999). It iswell established in South Africa and is being evalu-ated. However, it appears not to have established inMalawi (M. Hill, pers. comm. 2000). This insect hasbeen imported into Thailand for study but has not yetbeen released into the field.
Xubida infusellusBiology
Eggs are 0.52 mm by 0.87 mm long and are depos-ited in groups in an elongated gelatinous mass up toseveral centimetres long. Eggs hatch after 6 days.Larvae enter the laminae or petiole and tunnel down-wards, eventually entering the rhizome. There are 7–10instars and the development of larvae takes about 48days. The final instar larvae are about 25 mm long.
Table 4. Niphograpta albigutallis. Status of releases for each country. Data modified from Julien and Griffiths (1998)
Established Control
Year released
Unknown Under assessment
No Yes No Yes Under evaluation
Ghana 1996
Zimbabwe 1994
Panama 1977
Malawi 1996
Benin 1993
PNG 1994
Zambia 19711997a
a. Hill (1997).
South Africa 1990a
Sudan 1980
Australia 1977
USA 1977
Malaysia 1996
Thailand 1995
16
They move from the rhizome into a petiole where theytunnel to the surface. From the inside of the plant thelarvae cover the exit hole with a silken window andthen pupate in the tunnel below the window. The pupaeare about 20 mm long, do not have a cocoon andrequire 9 days to develop to adults. The adult emergesfrom the petiole through the window. Mating occurs onthe first night of emergence and most eggs are depos-ited in the second and third nights. Oviposition is notrestricted to water hyacinth and may occur on otherplant species, on pots and cage material. The number ofeggs masses per female (4–26) and number of eggs perfemale (180–684) are quite variable. Females live for4–8 days. Development from egg to adult is completedin about 64 days. This moth is susceptible to diseases,and variations in recorded biology may be due to vari-ations in the disease status of the colonies that werestudied. For greater detail see Silvera Guido (1971),DeLoach et al. (1980), and Sands and Kassulke (1983).
DamageThis moth attacks the older, slender petiole form of
the weed and should complement the damage caused bythe moth N. albigutallis, which prefers to attack young,tender plants, typified by the short, bulbous growthform. Young larvae of X. infusellus tunnelling insidethe petiole may girdle the petiole causing the portionabove the girdle to wilt and die. Feeding by largerlarvae in the lower petioles and rhizome severely debil-itates the plant and destroys apical meristems. Undercaged conditions, damage by larvae destroys plants.
ReleasesThis moth was first released in Australia in 1981
where it persisted for up to 13 months at two locationsbefore the demise of water hyacinth at those sites as aresult of human activity or drought. X. infusellus wasimported into Australia again for further study andreleased in Australia and Papua New Guinea (PNG)during 1996. Populations have persisted at one site inAustralia for over three years with no apparent impacton the plants. Adults were recorded at a release site inPNG on several occasions up to 18 months afterrelease, suggesting that the moth was established(Julien and Stanley 1999). However, no further assess-ments have been conducted in PNG since 1998.
Orthogalumna terebrantisBiology
Eggs are placed in small holes in the surface ofleaves and hatch in 7–8 days. The ensuing larvae areless than 0.24 mm long. Thereafter, three nymphal
stages occur, the final stage being up to 0.5 mm long.Development of larvae and nymphs requires 15 days.For details see Cordo and DeLoach (1975, 1976) andDel Fosse et al. (1975).
DamageThe nymphs of this sucking mite form galleries
between the parallel veins of the laminae from whichadults emerge. High populations of the mite cause leafdiscoloration and desiccation. Although this mite hasinfested various populations of water hyacinth for con-siderable periods it has not contributed to control ofthe weed.
Releases O. terebrantis was first released in Zambia in 1971
and in India during 1986. It is present in Mexico, Cuba,Jamaica, the southern USA and South America, andhas spread from Zambia to Malawi, Mozambique,South Africa and Zimbabwe (Julien and Griffiths1998). The impact of the mite, along with other agents,is being studied in Malawi (M. Hill, pers. comm. 2000).
Potential Agents Recently Considered or Currently Being
Studied
A resurgence of interest in better management of waterhyacinth, partially in response to the serious andincreasing water hyacinth problems in Africa, resultedin renewed interest in the studies of known potentialagents and the search for new agents. For example,recent studies were conducted and releases were madeof X. infusellus and E. catarinensis (see earlier). Otherinsect species have recently been assessed and rejectedas insufficiently host specific, while others are cur-rently being studied. They include the following.• The moth B. densa Walker has been rejected
because it attacks taro, Colocasia esculenta (L.)Schott. (Center and Hill 1999).
• The moth P. tenuis has been rejected as itdeveloped on a range of plants over severalfamilies (Cordo 1999).
• The grasshopper C. aquaticum is currently understudy in South Africa to clarify its host range(Oberholzer and Hill 2001).
• Thrypticus species flies are being studied inArgentina. Until recently this group of flies wasthought to have low priority because of suspectedwide host acceptance. Current studies have identifieda number of species within the group, one or moreapparently specific to water hyacinth (Cordo 1999).
17
Factors that Affect Establishment of Biological Control Agents
After identifying host-specific natural enemies that aresuitable for introduction and release, the next mostimportant step is establishing the agent(s) in the field.For those countries that release known agents, estab-lishing the agent is the first and most important step.Successful establishment is a prerequisite to control.The researcher can influence some issues that affectestablishment, and lack of attention to these can limitor prevent progress. They include: site selection,obtaining and maintaining healthy colonies of agentsfor mass rearing, rearing and releasing healthy andfecund individuals, and, depending on the dispersalcapacity of each agent, repeated and multiple releases.These issues are discussed by Wright (1997a,b) andJulien et al. (1999) for Neochetina species.
The Impact of Biological Control on Water Hyacinth Infestations
Of the six organisms that have been released, four (thetwo Neochetina weevils, the moth N. albigutallis andthe mite O. terebrantis) have been released eitherwidely or for long periods (Table 1). The two weevilshave provided excellent control in some habitats andhave contributed much less or not at all in others. It ismore difficult to assess the effects of the moth. Itsimpact on populations of the weed is insidious, andhard to quantify. It targets new, tender plants, typi-cally those on the edge of expanding mats, regrowthplants or those plants involved in early invasion. Thedamage caused by the moth is unlikely to controlserious infestations of the weed. However, it appearsto complement the actions of other control methods,both biological and non-biological, by reducingspread and invasiveness. The fourth organism, themite, has been established for about 30 years in Africaand USA. It has failed to contribute to control in itsown right, but there is conjecture that it may debilitatethe weed and therefore may contribute to control inthe presence of other factors.
Where biological control of water hyacinth isdemonstrably successful it has been a result of theactivities of either N. eichhorniae or N. bruchi or both.It is timely to assess the factors that have influencedthe successes and to try to identify the factors that haverestricted or prevent control occurring. Such informa-tion may help in future management of the weed. Itmay assist in making and testing predictions about
impact and control and in deciding how to developintegrated management strategies.
Successful biological control has occurred invarious locations in the following countries: Argen-tina, Australia, India, USA, PNG, Zimbabwe, the threeLake Victoria countries (Uganda, Tanzania, Kenya),South Africa and Thailand (Harley 1990; Julien et al.1999; Hill and Olckers 2001). The attributes of most ofthese locations are as follows.
• They are subtropical or tropical areas.• The weed mostly grew as a monoculture and not as
an understorey plant. • The weed was free to sink once damaged and was
not supported by other growth; nor were the rootsresting in mud beneath water.
• The mats were stable for long periods so that insectnumbers could build up.
• The weed was not subjected to regular removal byperiodic or annual flows and so insect densityincreased unabated to damaging levels.
• In some instances, the action of wind and wavesassisted the rate of damage and sinking of mats, e.g.Lake Victoria, Uganda. It is probable that controlwould have occurred regardless, although the levelof control may have been less. In other locations, thelack of the additional stresses on the damaged plantsimposed by wind and wave buffeting may limitcontrol (Hill and Olckers 2001).
• In other instances, the reduction in plant growth andstature, resulting from insect attack, caused mats todisintegrate into smaller components that could beflushed from lagoons via narrow channels andhence to the ocean, e.g. lagoon of the Sepik River.This flushing-out accelerated the rate of removal ofwater hyacinth from the system, but it is likely thatthe heavily damaged plants would have beendestroyed and sunk anyway, as occurred at otherimpounded locations, e.g. Lake Phayao, Thailandand Crescent Lagoon, Australia (A. Wright, pers.comm. 2000).High nutrient status of the plant may influence the
rate of control in tropical areas by allowing rapidincrease in insect populations. High nutrients maywork against control in temperate regions where theinsect activity is curtailed by cool winter conditions.As spring and summer approach, the weed is able torapidly outgrow previous damage before insect popu-lations have time to increase. Once insect populationsreach high proportions there is insufficient time in theremaining summer period to significantly damage theweed populations. Even if mat collapse occurs at the
18
end of the season, the seed would not have beendepleted (plants flower within six weeks of germina-tion) and reinvasion is inevitable. In such situations,appropriate intervention with other methods mightprovide control.
Disruption of biological control by inappropriateuse of herbicides is another reason for failure. Thismay occur where infestations are at important loca-tions and require immediate removal, in which casebiological control is inappropriate. It also occurs whenmanagers or politicians become frustrated waiting forbiological control to become evident. In many situa-tions, planned strategies could utilise biologicalcontrol to reduce the weed in the source area over thelong-term while shorter-term controls are used toreduce the problem at the critical points. When biolog-ical control becomes effective, three or more yearsafter release of the weevils, the water hyacinth at thesource will be reduced and hence the need to applyshort-term controls downstream should decline, e.g.Pangani River, Tanzania.
Conversely, it is important to identify the factorsthat may militate against control. These include thefollowing.
• Locations that experience temperate climates whereperiods of low temperature reduce or stop weevilpopulation increase and allow the weed to recover,e.g. areas in South Africa (Hill and Olckers 2001).
• High nutrient status of the water in temperateregions. See discussion above.
• Catastrophic reductions of the weevil populationsby periodic or annual floods. The weed populationscan recover much faster than the insect populationsand hence control is prevented.
• Catastrophic reductions in the weed biomass andinsect populations because of drought. The insectpopulations are driven to local extinction in theabsence of the host plant, whereas the waterhyacinth population continues with seedling growthafter rain (Hill and Olckers 2001).
• Sudd formation that prevents damaged waterhyacinth from sinking and provides a floatingreceptacle for seeds and a seedling bed.
• Shallow water where roots are embedded in mudand debris that may limit pupation, preventdamaged plants from sinking and encourage thegrowth of other plant species e.g. Melaleuca forestswampland in Australia (A. Wright, pers. comm.2000) and shallow inlets of Lake Kyoga, Uganda(J. Ogwang, pers. comm. 1999).
• The uptake of heavy metals by water hyacinth mayreduce fecundity of the weevils that feed on thoseplants (Jamil and Hussain 1993).
• Inappropriate application of other control methodsmay restrict the impact of biological control.Herbicide applications or physical removal mayeliminate establishing populations. They may limitincrease of established populations and reduceestablishing populations by killing plants thatsupport the insects. Application of some chemicalsto the weed may directly affect some control agents.
Interactions between the many environmentalfactors affect survivorship and population dynamics ofeach biological control agent and hence the level ofdamage and control. As a consequence, for eachcontrol agent, there is likely to be a range of controloutcomes, from areas where excellent control isachieved to those where biological control may have noimpact. For those locations where water hyacinth con-tinues to grow at greater than acceptable levels, man-agement should aim to make best use of the cheapestand most sustainable control method, normally biolog-ical control, in synergy with other available tools—herbicides, physical removal, manipulation of flows,and reductions of nutrient input.
Biological control is being developed through thesearch for new organisms to assist control in thoselocations where less than satisfactory control can beachieved with the current agents. There is room toimprove biological control and this is proceeding withthe research being conducted by USDA-ARS, ARC-PPRI and CABI Bioscience. However, it is unrealisticto think that biological control on its own will solve allwater hyacinth problems. Hence, there is a need todevelop integrated management strategies. Thismeans selecting the most appropriate control tech-niques available and implementing those techniquesso that they complement each other in time and space.The objective should be to obtain the best level ofcontrol that is affordable and sustainable while consid-ering environmental impacts. Since affordability andsustainability are major considerations in the manage-ment of most water hyacinth problems, biologicalcontrol should be the base component of all strategies.
References
Barrett, S.C.H. and Forno, I.W. 1982. Style morphdistribution in New World populations of Eichhorniacrassipes (Mart.) Solms-Laubach (water hyacinth).Aquatic Botany, 13, 299–306.
19
Bennett, F.D. and Zwölfer, H. 1968. Exploration for naturalenemies of the water hyacinth in northern South Americaand Trinidad. Hyacinth Control Journal, 7, 44–52.
Center, T.D. 1981. Release and establishment of Sameodesalbiguttalis for biological control of waterhyacinth. USArmy Engineer Waterways Experiment Station,Vicksburg, Mississippi. Technical Report A-81-3, 120p.
Center, T.D. 1994. Biological control of weeds:waterhyacinth and waterlettuce, In: Rosen, D., Bennett,F.D. and Capinera, J.L. ed., Pest management in thesubtropics. biological control—a Florida perspective.UK, Intercept Ltd, 481–521.
Center, T.D. and Hill, M.P. 1999. Host specificity of thepickerelweed borer, Bellura densa Walker (Lepidoptera:Noctuidae) a potentially damaging natural enemy ofwater hyacinth. In: Hill, M.P., Julien, M.H. and Center,T.D., ed., Proceeding of the 1st IOBC Water HyacinthWorking Group, Harare, November 1998, 67. (abstract)
Charudattan, R. 2001. Control of water hyacinth by the useof plant pathogens. These proceedings.
Cordo, H.A. 1999. New agents for biological control ofwater hyacinth. In: Hill, M.P., Julien, M.H. and Center,T.D., ed., Proceedings of the 1st IOBC Water HyacinthWorking Group, Harare, November 1998, 68–74.
Cordo, H.A. and DeLoach, C.J. 1975. Ovipositionalspecificity and feeding habits of the waqterhyacinth mite,Orthogalumna terebrantis, in Argentina. EnvironmentalEntomology, 4, 561–565.
Cordo, H.A. and DeLoach, C.J. 1976. Biology of thewaterhyacinth mite in Argentina. Weed Science, 24, 245–249.
Del Fosse, E.S., Cromroy, H.L. and Habeck, D.H. 1975.Determinations of the feeding mechanisms of thewaterhyacinth mite. Hyacinth Control Journal, 13, 53–55.
DeLoach, C.J. and Cordo, H.A. 1978. Life history andecology of the moth Sameodes albiguttalis, a candidatefor biological control of waterhyacinth. EnvironmentalEntomology, 7, 309–321.
DeLoach, C.J., Cordo, H.A., Ferrer, R. and Runnacles,J.1980. Acigona infusella, a potential biological controlagent for waterhyacinth: observations in Argentina (withdescriptions of two new Apanteles by L. De Santis).Annals of the Entomological Society of America, 73,138–146.
Ding, J., Wang, R., Fu, W. and Zhang, G. 2001. Waterhyacinth in China: its distribution, problems, and controlstatus. These proceedings.
Fayad, Y.H., Amira, A.I., Amal, A.E. and Fawzy, F.S. 2001.Ongoing activities on the biological control of waterhyacinth in Egypt. These proceedings.
Harley, K.L.S. 1990. The role of biological control in themanagement of water hyacinth, Eichhornia crassipes.Biocontrol News and Information, 11, 11–22.
Harley, K.L.S., Julien, M.H. and Wright, A.D. 1996. Waterhyacinth: a tropical worldwide problem and methods forits control. Proceedings of the 2nd International WeedControl Congress, Copenhagen, 1996, volume II, 639–644.
Heard, T.A. and Winterton, S.L. 2000. Interactions betweennutrient status and weevil herbivory in the biologicalcontrol of water hyacinth. Journal of Applied Ecology,37, 117–127.
Hill, M.P. 1997. Water hyacinth in Zambia: restoring thebalance on the Kafue River. Plant Protection News, 47,11–13.
Hill, M.P., Cilliers, C.J. and Neser, S. 1999. Life history andlaboratory host range of Eccritotarsus catarinensis(Carvahlo) (Heteroptera: Miridae), a new natural enemyreleased on waterhyacinth (Eichhorniae crassipes (Mart.)Solms-Laub) (Pontederiaceae) in South Africa.Biological Control, 14, 127–133.
Hill, M.P. and Olckers, T. 2001. Biological controlinitiatives against water hyacinth in South Africa:constraining factors, success and new courses of action.These proceedings.
IITA (International Institute for Tropical Agriculture) 2000.Annual Report of the IITA for 2000, Cotonou, Benin, 20.
Jamil, K. and Hussain, S. 1993. Biochemical variations inovaries of water hyacinth weevils Neochetinaeichhorniae (Warner) exposed to heavy metals. IndianJournal of Experimental Biology, 32, 36–40.
Julien, M.H. and Griffiths, M.W. 1998. Biological control ofweeds. A world catalogue of agents and their targetweeds, 4th Edition. CABI Publishing, Wallingford, 223p.
Julien, M.H., Griffiths, M.W. and Wright, A.D. 1999.Biological control of water hyacinth. The weevilsNeochetina bruchi and Neochetina eichhorniae:biologies, host ranges, and rearing, releasing andmonitoring techniques for biological control ofEichhornia crassipes. ACIAR Monograph No. 60, 87p.
Julien, M.H., Harley, K.L.S., Wright, A.D., Cilliers, C., Hill,M., Gunter, T., Cordo, H. and Cofrancesco, A. 1996.International cooperation and linkages in themanagement of water hyacinth with emphasis onbiological control. In: Moran, V.C. and Hoffman, J.H.,ed., Proceedings of the 9th International Symposium onBiological Control of Weeds, Stellenbosch, South Africa1996, 273–282.
Julien, M.H. and Stanley, J.N. 1999. Recent research onbiological control for water hyacinth in Australia. In: Hill,M.P., Julien, M.H. and Center, T.D., ed., Proceedings ofthe 1st IOBC Water Hyacinth Working Group, Harare,November 1998, 52–61.
Moorehouse, T., Agaba, P. and McNabb, T. 2001.Biological control of water hyacinth in the Kagera Riverheadwaters of Rwanda. These proceedings.
20
Oberholzer, I.G. and Hill, M.P. 2001. How safe is thegrasshopper Cornops aquaticum for release on waterhyacinth in South Africa? These proceedings.
Sands, D.P.A. and Kassulke, R.C. 1983. Acigona infusella(Walker) (Lepidoptera: Pyralidae), an agent forbiological control of water hyacinth (Eichhorniacrassipes) in Australia. Bulletin of EntomologicalResearch, 73, 625–632.
Silvera Guido, A. 1971. Preliminary data on biology andspecificity of Acigona ignitalis Hamps. (Lep. Pyralidae)on the host plant Eichhornia crassipes (Mart.) Solms-Laubach (Pontederiaceae). Revista de la SociedadEntomologia Argentina, 33, 137–145. (In Spanish).
Stanley, J.N. and Julien, M.H. 1999. The host range ofEccritotarsus catarinensis (Heterophera: Miridae) apotential agent for the biological control of waterhyacinth(Eichhornia crassipes). Biological Control, 14, 400–404.
Wright, A.D. 1997a. Distribution of agents. In: Julien,M.H. and White, G. ed. Biological control of weeds:theory and practical application. ACIAR MonographsNo. 49, 97–100.
Wright, A.D. 1997b. Establishment of agents. In: Julien,M.H. and White, G. ed. Biological control of weeds:theory and practical application. ACIAR Monograph No.49, 101–103.
21
Biological Control of Water Hyacinth by Using Pathogens: Opportunities, Challenges, and Recent
Developments
R. Charudattan*
Abstract
There is good justification to renew concerted efforts to develop pathogens for biological control of waterhyacinth (Eichhornia crassipes, (Mart.) Solms-Laub.). Among the most promising pathogens are Uredoeichhorniae, suitable as a classical biocontrol agent, and Acremonium zonatum, Alternaria eichhorniae,Cercospora piaropi, Myrothecium roridum, and Rhizoctonia solani, which are widely distributed in differentcontinents, as bioherbicides. Other, less widely distributed pathogens, notably species of Bipolaris,Drechslera, and Fusarium, may hold promise, but further studies are needed to confirm their usefulness.Ongoing studies on the biology of U. eichhorniae are expected to help resolve the biocontrol potential of thisrust fungus in the near future. It is also anticipated that recent advances in bioherbicide technology, andprevious experience with pathogens of water hyacinth, will enable development of effective bioherbicides.Success in this effort will require the use of highly virulent pathogens and pathogen strains as well as novelformulations that help to counter water hyacinth’s ability for rapid growth under different site conditions andassist the pathogen to overcome environmental limitations. In Florida, USA, our aim is to develop abioherbicide that could be used in integration with the existing suite of introduced arthropod biocontrol agentsand improve the overall effectiveness of the biological control strategy under different weed-control scenarios.In this attempt, unlike in previous unsuccessful attempts, we are aided by the availability of a large and diversecollection of highly virulent pathogen strains to choose from, the facility to formulate effective pathogens innewer materials to assure consistency of performance, the ability to integrate dual and multiple pathogens, anda ‘biofriendly’ posture of regulatory agencies toward the development and registration of bioherbicides.Results from our ongoing field trials suggest the feasibility and commercial potential of our current approach.
THERE are several good reasons to consider pathogensas biocontrol agents: pathogens can cause significantreductions in water hyacinth biomass, especially fol-lowing natural disease outbreaks, after severe insectattacks, or when used as inundative bioherbicideagents (Charudattan et al. 1985; Shabana et al. 1995b).Published accounts of natural declines in water hya-cinth populations following natural disease outbreaks
further confirm the potential of pathogens in limitingwater hyacinth populations (Martyn 1985; Morris1990). Controlled experimental studies have con-firmed the potential of Acremonium zonatum, Alter-naria eichhorniae, and Cercospora piaropi to control(i.e. reduce weed biomass) water hyacinth (Martynand Freeman 1978; Charudattan et al. 1985; Shabanaet al. 1995b). In addition, it has been well proven thatpathogens can be successful as classical or inundative(bioherbicide) agents. Currently, worldwide about 15–20 weeds are biologically controlled with pathogens(Rosskopf et al. 1999). Therefore, there is clear justi-
* Center for Aquatic and Invasive Plants, Plant PathologyDepartment, University of Florida, Gainesville, FL 32611-0680, USA. Email: [email protected]
22
fication to explore pathogens fully and fairly as agentsof biological control of water hyacinth. In this effort,we will be starting with the benefit of past attemptsthat have helped to lay the foundation through surveysand characterisation of host–pathogen systems. It isnow imperative to take steps to turn this early empir-ical knowledge and success into practical utilisation ofpathogens in weed-management programs. Clearly,with some pathogens, such as the rust fungus Uredoeichhorniae, further studies are needed to understandthe biology and biocontrol potential of the agentsbefore they can be used. Nonetheless, the current stateof knowledge and experience give us the optimism toassert that pathogens can succeed in practice, espe-cially when used in integrated weed control systemsrather than as stand-alone control options.
Opportunities and Challenges
Development of a pathogen or pathogens for use inoperational weed management systems is longoverdue. In this regard, the present timing and oppor-tunities may be the best we can ever expect. Forinstance, there is renewed interest in developing addi-tional biological control agents to supplement thosealready in use. There is an urgent need for effectivecontrols for water hyacinth in countries of Africa andAsia. Support, in the form of funding initiatives, is alsoevident, as in the case of the recently formed Pan-African mycoherbicide program. Despite these opti-mistic signs, the most important lesson learned fromearlier attempts should not be forgotten: that foliarpathogens generally do not have the capacity to killwater hyacinth plants completely and quickly unlessthey can be used in conjunction with efficacy-enhancing formulations and adjuvants, low doses ofchemical synergists, and/or insect biocontrol agents.Bioherbicides were regarded as biological substitutesfor chemical herbicides and therefore as stand-aloneproducts. There was little technological sophisticationin the bioherbicide products; usually fresh fungal inoc-ulum was used without formulations devised to protectthe inoculum from adverse environmental conditionsor to improve its performance. Improvements in effi-cacy of foliar pathogens are now possible through anumber of recently developed formulations such ashydrophilic polymers, emulsions, surfactants etc.(Shabana et al. 1997; Green et al. 1998; Boyetchko etal. 1999). Several pathogens can be combined andused in a ‘multiple-pathogen strategy’ (Chan-dramohan et al. 2000) to improve the level of weedcontrol, minimise or prevent development of host
resistance, overcome age-related host resistance,assure consistency in bioherbicide performance,improve environmental latitude of activity, and so on.It is also possible to ‘customise’ the pathogen mixturedepending on the types of pathogens available for usein a given country or region.
Among the challenges, we must address the perva-sive perception among aquatic weed managers thatbiological control does not provide quick and accept-able levels of control and that biological control isinadequate under conditions that promote rapid expan-sion of water hyacinth mats. Given the general con-sensus of this working group that the overalleffectiveness of biological control should be (couldbe) improved, we should aim to develop a pathogen orpathogens that can provide quick and assured levels ofbiological control. Accordingly, we should explore allpotentially useful pathogens. In this respect, both clas-sical and bioherbicide strategies should be pursued.
As we renew our efforts to develop pathogens, weshould remember the following.• Localised pathogens, fungi and bacteria that have not
been previously explored, may be useful. Therefore,further surveys, especially in the African continentand the Neotropics, and thorough evaluations of thepathogens found, should be a part of our plans.
• Technological innovations should be developed toovercome micro-environmental conditionssurrounding the plant that are non-conducive fordisease development.
• Water hyacinth plants growing under conditionsthat promote rapid growth will result in the plantoutgrowing the rate of progress of foliar diseases.Consequently, soon after a bioherbicide application,the disease pressure will wane, turning even adisease with polycyclic potential into a monocyclicdisease. Therefore, the ability of water hyacinthplants to grow at rapid rates under different siteconditions is a challenge that must be factored intothe design of bioherbicides.To address the dual challenges of rapid host growth
rate and environmental constraints and to assure bio-control effectiveness, the pathogen (specifically a bio-herbicide) may be applied with low rates of aregistered chemical herbicide. Alternatively, the bio-herbicide formulation may contain an adjuvant (e.g. aphytotoxic compound or a registered chemical herbi-cide at low rates). Multiple applications of the bioher-bicide may be also used, but the economics of multipleapplications would have to be assessed. Also, bioher-bicide applications may be timed to maximise theimpacts of insect agents and thus increase the level of
23
biotic stress. Novel formulations should be developedthat help to prolong humidity on the leaf surface,protect the pathogen against solar irradiation, and/orpromote leaf penetration by the pathogen. Combina-tions of two or three different pathogens may be usedto increase the level of damage and consistency of per-formance, as has been tried under experimental condi-tions, most recently by Den Breeÿen (1999) andVincent and Charudattan (2000). An approach that isquite applicable but has not been tried is the use ofseveral strains of a pathogen (e.g. C. piaropi), eachhaving different levels of virulence, fitness, phyto-toxin production, and other desirable traits.
It is fair to say that, despite several field surveys in thepast two decades, no new and highly promising patho-gens have been added to the list of known, prospectivebiocontrol agents. However, this statement is not meantto imply that there may be no additional pathogen can-didates left to discover, but rather it is an assessment ofthe current situation. For example, of the nearly 70 fungiand bacteria recorded on water hyacinth (Barreto andEvans 1996; Charudattan 1996), only about 15 havebeen adequately tested and confirmed to be highly vir-ulent pathogens. Of these, three fungal pathogens, Acre-monium zonatum, Alternaria eichhorniae, andCercospora piaropi (= C. rodmanii), have been studiedintensively as biocontrol agents and shown to be effec-tive in controlling water hyacinth under experimentalconditions (Martyn and Freeman 1978; Charudattan etal. 1985; Shabana et al. 1995b). This leaves a largenumber of other reported fungi and bacteria that remainto be assessed for their biocontrol potential. Thus, fornow, the choice of pathogens for biological control islimited to Uredo eichhorniae, as a classical biocontrolagent, and as bioherbicide agents, A. zonatum, A. eich-horniae, C. piaropi (= C. rodmanii), Myrotheciumroridum, and Rhizoctonia solani. Several other, lesswidely distributed pathogens, such as species of Bipo-laris, Drechslera, and Fusarium, may hold promise, butthey need to be studied further to confirm their potential.
A Synopsis of Highly Virulent and Useful Pathogens
The following are brief descriptions of virulentpathogens that are leading candidates for furtherdevelopment.
Acremonium zonatum
This fungus causes an easily identified necroticzonate leaf spot characterised by spreading lesions,
most noticeable on the upper laminar surface (Fig. 1).On the lower surface, which is normally protectedfrom direct sunlight, the area directly under the spotmay have a sparse, spreading layer of white fungal(mycelial) growth. Each spot may be small (2 mmdiameter) to large (> 3 cm diameter) and the spots maycoalesce, covering most of the lamina. The zonatepattern may not be evident in new infections whenmost spots are small. This disease has been reportedfrom Australia, USA, and many countries of Asia,Central America, and South America. It is often asso-ciated in the field with infestations of the water hya-cinth mite Orthogalumna terebrantis. This pathogen isrepresented by several highly virulent strains such asthe ones found in Mexico by Martinez Jimenez andCharudattan (1998).
Alternaria eichhorniae
Two species of Alternaria, A. eichhorniae and A.alternata, have been recorded on water hyacinth. Oneor both of these species have been reported from Aus-tralia, Bangladesh, Egypt, India, Indonesia, and SouthAfrica. Alternaria alternata appears to be a weak,opportunistic parasite, whereas A. eichhorniae is ahighly virulent, host-specific pathogen of water hya-cinth. Alternaria eichhorniae has been shown to havegood potential as a bioherbicide agent (Shabana et al.1995a,b; these proceedings). It causes discretenecrotic foliar spots (oblong, 2–4 mm long) sur-rounded by a bright yellow halo. Blighting of theentire leaf lamina can be induced by using mycelialinoculum and providing prolonged, 100% relativehumidity (Fig. 2). In culture, A. eichhorniae producesseveral bright red compounds in culture, includingbostrycin and deoxybostrycin that are phytotoxic towater hyacinth leaves (Charudattan and Rao 1982).The extent of naturally occurring variability in viru-lence in this pathogen is not clear. More details aboutthis pathogen are given by Shabana (2001).
Cercospora piaropi (= C. rodmanii)
Symptoms caused by Cercospora spp. may be easilyconfused with those of many other foliar pathogens,including many opportunistic, weak parasites. Untilnow, two species of Cercospora, C. piaropi and C. rod-manii, have been recognised as pathogens of waterhyacinth, but recently Tessmann et al. (2001) havemerged the two species into an emended C. piaropi.Cercospora piaropi has been reported on water hya-cinth from throughout the present range of the weed.This pathogen causes small (2–4 mm diameter)
24
necrotic spots on laminae and petioles (Fig. 3). Thespots are characterised by pale centres surrounded bydarker necrotic regions. Occasionally, the spots mayappear in the shape of ‘teardrops’ that coalesce as theleaf matures, causing the entire leaf to turn necrotic andsenescent. In fact, the senescence is accelerated by theCercospora disease, and the disease can rapidly spreadacross water hyacinth infestations, causing large areasof the weed mat to turn brown and necrotic. Undersevere infections, the plant may be physiologicallystressed, lose its ability to regenerate, become water-logged, and sink or disintegrate.
Tessmann et al. (2000) compared 60 isolates of Cer-cospora species isolated from water hyacinth leavesshowing symptoms of Cercospora infection whichwere collected from the USA, Mexico, Venezuela,Brazil, South Africa, and Zambia. They found the iso-lates to be variable in pigmentation in culture, sporemorphology, and virulence. Virulence of the isolateswas also variable: isolates ranged from being nearlyavirulent to highly virulent and capable of causing leafdeath. These traits were independent of the geographicorigin of isolates. The isolates were then tested to seeif the species concept based on conidial and culturalmorphology and virulence, as used by Conway (1976)in his designation of C. rodmanii into a separate spe-cies, might agree with a species concept developedwith the help of molecular markers.
Accordingly, members of a collection of isolatesrepresenting acquisitions from the USA (Florida andTexas), Mexico, Venezuela, Brazil, South Africa, andZambia were compared on the basis of the DNAsequences of gene segments for beta-tubulin (TUB2),histone-3 (H3), and elongation factor-1-alpha (EF1a),corresponding to 380, 309, and 431 base pairs (bp),respectively. Eight of the isolates were also comparedfor the rDNA regions containing ITS1, ITS2, and the5.8S gene. Extracted DNA was amplified bypolymerase chain reaction using TUB2, H3, and ITSprimer pairs selected from the literature. The com-bined phylogenetic relationships of TUB2, H3, andEF1a sequences done with phylogenetic analysisusing parsimony did not support the species distinc-tion between C. piaropi and C. rodmanii. Isolates rep-resentative of both species grouped into a single, well-supported clade (Tessmann et al. 2000).
Thus, the molecular evidence pointed strongly to acommon phylogeny of Cercospora pathogenic towater hyacinth. This raises some important questionsrelevant to the use of C. piaropi for water hyacinthcontrol. For instance, given the worldwide distributionof this pathogen and the molecular evidence pointing
to a common origin of Cercospora isolates pathogenicto water hyacinth, should this pathogen be subject toplant quarantine restrictions? Would it not be advanta-geous to import and supplement native C. piaropistrains with more highly virulent strains from which-ever continent or country they could be found? Perhapsa search for highly virulent strains (of any pathogen)should be an inherent priority for all mycoherbicideprograms in order to ensure that only the best strainsare used. These questions deserve to be considered.
Myrothecium roridum
This fungus causes a teardrop-shaped leaf spot (up to1 × 5 cm), rounded on the side facing the petiole andtapering to a narrow point in the direction of thelaminar tip (Fig. 4). Older leaf spots turn necrotic withdark brown margins, with the centre of the spot coveredwith discrete white and black conidial masses.Myrothecium disease of water hyacinth has beenreported to occur in India, Malaysia, Indonesia, pos-sibly Mexico, and some western African countries.Although this species is worldwide in distribution, thetypical myrothecium disease has not been recorded onwater hyacinth in the Americas. The occurrence of var-iability in virulence in this pathogen is therefore notclear. Some recent studies suggest that someMyrothecium species can be used as broad-spectrumbioherbicides against several weeds (Walker and Tilley1997), a finding that has implications for the develop-ment of M. roridum for water hyacinth control.
Rhizoctonia solani
Disease symptoms caused by this fungus mayresemble damage caused by a desiccant type of chem-ical herbicide (e.g. diquat). Symptoms consist of irreg-ular, necrotic spots, and broad lesions (Fig. 5). Unlikechemical damage, the brown necrotic areas are usuallysurrounded by noticeable, thin, water-soaked marginsof darker brown colour than the rest of the lesion.Rhizoctonia disease has been reported on water hya-cinth from the southeastern United States, Brazil,Mexico, Panama, Puerto Rico, India, Malaysia, andIndonesia. This fungus is usually very aggressive anddestructive, capable of rapidly killing water hyacinthplants. The extent of variability in virulence of R.solani pathogenic to water hyacinth is not clear, butisolates collected in the USA, Panama, and Brazil havebeen found to be extremely virulent (R. Charudattan,unpublished data; R.A. Pitelli, University of the Stateof Sao Paulo, Jaboticabal, Brazil, pers. comm.).
25
Figure 4. Teardrop-shaped foliar lesions and the extentof damage caused by Myrothecium roridum
Figure 5. Blighting symptoms caused by Rhizoctoniasolani on leaves (left) and whole plants(right). Picture on the right shows fungus-infected and uninfected control plants
Figure 6. Water hyacinth leaves showing uredialpustules of Uredo eichhorniae
Figure 3. Leaf spot symptoms caused Cercosporapiaropi
Figure 2. Discrete necrotic leaf spots surrounded byyellow halo (left), symptoms developedwhen inoculated with spores (middle) ormycelium and kept under high humidity(right)
Figure 1. Zonate leaf spots caused by Acremoniumzonatum (left) and naturally infected plantsin the field (right)
26
Despite its high virulence and destructive capabili-ties, R. solani has never been seriously considered as abioherbicide agent because of its reputed wide hostrange. However, this should not be a deterrent intoday’s regulatory environment which has permittedregistration and commercial use of pathogens that donot possess high levels of host specificity (e.g. Colle-totrichum gloeosporioides f.sp. aeschynomene,Phytophthora palmivora, and Chondrostereum pur-pureum).
Uredo eichhorniae
This rust fungus occurs in southern Brazil, Argen-tina, and Uruguay. It is known only in its uredial sporestage (Fig. 6). Because it is a rust fungus, it is likely tobe highly host-specific and therefore a desirable clas-sical biological control agent, but several aspects of itsbiology remain to be fully understood. For instance,since this fungus was first described from the Domin-ican Republic, a tropical island, it is unclear why itdoes not occur beyond its present range of distributionin the subtropical to temperate regions of SouthAmerica. It is possible that it is adapted to slowergrowing plants of the temperate fringes of water hya-cinth’s distribution and its original finding in theDominican Republic is an anomaly. Nevertheless, theoccurrence of this pathogen only inside the nativerange of water hyacinth, but not outside (e.g. Asia,Africa, or Australia), is to be expected on the basis ofecological theory of coevolution of rust fungi.
Our earlier attempt to import U. eichhorniae intoFlorida was disallowed by U.S. regulatory agencieson the grounds that the full life cycle of this funguswas unknown. Stimulated by the current interest indeploying additional biocontrol agents, we haverecently restarted research on the life cycle andbiology of this fungus in collaboration with scientistsfrom the Plant Protection Research Institute, Stellen-bosch, South Africa and the University of the State ofSao Paulo at Jaboticabal, Brazil. The objective is toimport the rust into quarantine at two locationsoutside its native range, one in the southern hemi-sphere (Stellenbosch, South Africa), the other in thenorthern hemisphere (Gainesville, Florida), to initiateepidemio-logical studies under controlled conditions.In this regard, South Africa, with its climatic, latitu-dinal, and hemispheric similarity to temperate SouthAmerica, offers an eminently suitable location toconduct these studies.
As a first step in the evaluation of its biocontrolpotential, we have initiated studies on the life cycle
(i.e. spore stages, particularly the aeciospores and tel-iospores) and disease cycle of U. eichhorniae underfield conditions. The biocontrol potential of thisfungus will be determined by using fungicide(s) toblock the effects of the rust disease in natural fieldpopulations of water hyacinth and comparing thegrowth rates of such fungicide-protected plants withrust-infected plants. These studies should help toestablish the suitability of U. eichhorniae as a clas-sical biocontrol agent and set the stage for its eventualrelease into the USA, South Africa, and elsewhere.The addition of this rust pathogen to the existing suiteof biocontrol agents is likely to improve the prospectsfor a sustainable, long-term biological control ofwater hyacinth.
Recent Developments in Progress
With the above-mentioned concepts in mind, ourcurrent efforts in Florida are aimed at the develop-ment of a bioherbicide that can be used in combina-tion with existing insect biocontrol agents. Our goal isa bioherbicide that will help to improve the overalleffectiveness of the biological control system underdifferent control scenarios. Our emphasis is on aknock back (reduce biomass) rather than a knockdown (weed kill and biomass elimination) strategy.Specifically, we are evaluating two pathogens, C.piaropi and an isolate of M. roridum (isolated frombegonia since no virulent isolate of this pathogen hasbeen found on water hyacinth in the United States).The pathogens are being tested individually and incombination and applied with a surfactant, an invertemulsion, and/or a humectant gel. Applications aremade to water hyacinth plants with and withoutnatural populations of Neochetina spp. Results fromfield trials in progress indicate that a treatment with C.piaropi applied in the surfactant Silwet L-77 providedthe best levels of biomass reduction and damageseverity. The combination of the two pathogens wasnot significantly better than C. piaropi plus Silwet L-77, possibly because the isolate of M. roridum used isnot a true pathogen of water hyacinth (Vincent andCharudattan 2000). Further, large-scale field trials areunder way to develop a commercially feasible bioher-bicide formulation.
Conclusions
It is a challenge to develop an effective bioherbicidethat is acceptable for use in practical water hyacinthmanagement programs. The challenge can be met,
27
especially now, given the bio-friendly regulatoryclimate in the United States and other countries, andthe availability of newer, innovative approaches tobioherbicide development and deployment. An assess-ment of the potential usefulness of Uredo eichhorniaeas a classical biocontrol agent is in progress and theprospects for using this rust fungus on a global scaleshould be known in the near future.
Acknowledgments
I thank Dr Ding Jianqing and the Chinese Academyof Agricultural Sciences for their invitation and spon-sorship of my participation in this workshop. I extendmy appreciation to my graduate student Ms AngelaVincent for providing slides and data from herongoing field trials. Other students whose works Ihave reviewed in this presentation are Dr YasserShabana and Dr Dauri Tessmann. Our work onUredo eichhorniae has been supported by funds pro-vided by the US Department of Agriculture (USDA)-Agricultural Research Service, under the Coopera-tive Agreement #58-6629-0-203, and USDA-ForeignAgricultural Services grant #58-3148-7-015.
References
Barreto, R.W. and Evans, H.C. 1996. Fungal pathogens ofsome Brazilian aquatic weeds and their potential use inbiocontrol. In: Moran, V.C. and Hoffmann, J.H., ed.,Proceedings of the IX International Symposium onBiological Control of Weeds, Stellenbosch, South Africa,19–26 January, 1996. Cape Town, University of CapeTown, 121–126.
Boyetchko, S., Pedersen, E., Punja, Z. and Reddy, M. 1999.Formulations of biopesticides. In: Hall, F.R. and Menn,J.J., ed., Methods in biotechnology. Biopesticides: useand delivery. Totowa, New Jersey, Humana Press, 5,487–508.
Chandramohan, S., Charudattan, R., Megh Singh, andSonoda, R.M. 2000. Multiple pathogen strategy: a novelapproach for bioherbicidal control of several weeds. In:Anon., ed., Abstracts, Third International Weed ScienceCongress—IWSC, Foz do Iguassu, Brazil, 6–11 June2000. Corvallis, Oregon State University, 182–183.
Charudattan, R. 1996. Pathogens for biological control ofwater hyacinth. In: Charudattan, R., Labrada, R., Center,T.D. and Kelly-Begazo, C., ed., Strategies for waterhyacinth control, report of a panel of experts meeting,Fort Lauderdale, Florida, 11–14 September 1995.Gainesville, Florida, Institute of Food and AgriculturalSciences, University of Florida, 189–199.
Charudattan, R., Linda, S.B., Kluepfel, M., and Osman,Y.A. 1985. Biocontrol efficacy of Cercospora rodmaniion waterhyacinth. Phytopathology, 75, 1263–1269.
Charudattan, R. and Rao, K.V. 1982. Bostrycin and4-deoxybostrycin: two nonspecific phytotoxins producedby Alternaria eichhorniae. Applied EnvironmentalMicrobiology, 43, 846–849.
Conway, K. E. 1976. Cercospora rodmanii, a new pathogenof waterhyacinth with biological control potential.Canadian Journal of Botany, 54, 1079–83.
Den Breeÿen, A. 1999. Biological control of water hyacinthusing plant pathogens: dual pathogenicity and insectinteractions. In: Hill, M.P., Julien, M.H. and Center, T.D.,ed., Proceedings of the First IOBC Global WorkingGroup Meeting for the Biological and Integrated Controlof Water Hyacinth, Harare, Zimbabwe, 16–19 November1998. Pretoria, South Africa, Plant Protection ResearchInstitute, 75–79.
Green, S., Stewart-Wade, S.M., Boland, G.J., Teshler, M.P.and Liu, S.H. 1998. Formulating microorganisms forbiological control of weeds, In: Boland, G.J. andKuykendall, L.D., ed., Plant–microbe interactions andbiological control. New York, Marcel Dekker, 249–281.
Martinez Jimenez, M. and Charudattan, R. 1998. Survey andevaluation of Mexican native fungi for biocontrol ofwaterhyacinth. Journal of Aquatic Plant Management, 36,145–148.
Martyn, R.D. (1985). Waterhyacinth decline in Texascaused by Cercospora piaropi. Journal of Aquatic PlantManagement, 23, 29–32.
Martyn, R.D. and Freeman, T.E. 1978. Evaluation ofAcremonium zonatum as a potential biocontrol agent ofwaterhyacinth. Plant Disease Reports, 62, 604–608.
Morris, M.J. 1990. Cercospora piaropi recorded on theaquatic weed, Eichhornia crassipes, in South Africa.Phytophylactica, 22, 255–256.
Rosskopf, E.N., Charudattan, R. and Kadir, J.B. 1999. Use ofplant pathogens in weed control. In: Fisher, T.W.,Bellows, T.S., Caltagirone, L.E., Dahlsten, D.L.,Huffaker, C. and Gordh, G., ed., Handbook of biologicalcontrol. San Diego, California, Academic Press, 891–918.
Shabana, Y.M., Charudattan, R., DeValerio, J.T. andElwakil, M.A. 1997. An evaluation of hydrophilicpolymers for formulating the bioherbicide agentsAlternaria cassiae and A. eichhorniae. WeedTechnology, 11, 212–220.
Shabana, Y.M., Charudattan, R. and ElWakil, M.A. 1995a.Identification, pathogenicity, and safety of Alternariaeichhorniae from Egypt as a bioherbicide agent forwaterhyacinth. Biological Control, 5, 123–135.
— 1995b. Evaluation of Alternaria eichhorniae as abioherbicide for waterhyacinth (Eichhornia crassipes) ingreenhouse trials. Biological Control, 5,136–144.
28
Shabana, Y.M., Elwakil, M.A. and Charudattan, R. 2001.Biological control of water hyacinth by mycoherbicide inEygpt. These proceedings
Tessmann, D.J., Charudattan, R., Kistler, H.C. andRosskopf, E.N. 2000. A molecular characterization ofCercospora species pathogenic to water hyacinth andemendation of C. piaropi. Mycologia, in press.
Vincent, A.C. and Charudattan, R. 2000. Evaluation of acombination of two pathogens as a potential bioherbicide
for Eichhornia crassipes [Mart.] Solms-Laub. under fieldconditions. Abstracts, Third International Weed ScienceCongress—IWSC, Foz do Iguassu, Brazil, 6–11 June2000. Foz do Iguassu, Brazil, 217.
Walker, H.L. and Tilley, A.M. 1997. Evaluation of an isolateof Myrothecium verrucaria from sicklepod (Sennaobtusifolia) as a potential mycoherbicide agent.Biological Control, 10, 104–112.
29
Water Hyacinth in China: Its Distribution, Problems and Control Status
Ding Jianqing, Wang Ren, Fu Weidong and Zhang Guoliang*
Abstract
Water hyacinth is one of the most important invasive alien plant species in China, into which it was introducedin the early 1900s. Now the weed is distributed in 17 provinces, and in Guangdong, Yunnan, Fujian, Zhejiangand Taiwan has become a bio-disaster. Compared with the situation in 1960, few people now use waterhyacinth plants to feed pigs or ducks, or to make fertilisers. It poses a great threat to agriculture, fisheries,transportation and the environment. It is estimated that each year more than 100 million RMB yuan (US$12m)is spent on control of water hyacinth throughout China, but in most areas the weed remains vigorous andcontinues to spread. Chemical, mechanical and biological control, as well as integrated control strategies, havebeen employed to combat water hyacinth in more than 10 provinces. Two weevils, Neochetina eichhorniae andN. bruchi, which were introduced from Argentina and USA in 1995, have established and spread theirpopulations in Zhejiang and Fujian provinces. The weevils greatly suppressed the plants around the releaseareas. In early 2000, a mirid, Eccritotarsus catarinensis, was introduced from South Africa but did notestablish. A survey for pathogens of water hyacinth began in spring 2000 in southern China.
WATER hyacinth was introduced into China in the early1900s. As an ornamental plant, it was first introducedinto Taiwan in 1903 from Southeast Asia. In the 1930sit was introduced to the mainland (Diao 1989). But thefirst scientific record appeared for the mainland in 1954in the book, ‘Taxonomy Catalogue for China’s Plants:Families and Genera’ (Anon. 1954). In the 1950s and1960s, water hyacinth was distributed widely intoalmost all provinces for animal food. After artificialtransplanting and mass rearing and breeding, water hya-cinth was distributed to further areas in the 1970s andbegan to cause damage in the 1980s. Increasing damagehas been reported since the 1990s as nutrient levelsincreased in water bodies and the use of water hyacinthplants began to fall.
Distribution
Water hyacinth is now distributed naturally in 17 prov-inces or cities in China. In several other provinceswater hyacinth is still utilised but cannot overwinter.Water hyacinth causes damage in more than 10 prov-inces (Ding et al. 1995). Great damage has beenreported in five provinces: Yunnan, Guangdong, Zhe-jiang, Fujian, Taiwan. Figure 1 maps the distributionof water hyacinth in China.
Problems Arising from Water Hyacinth
As in many other countries, in China water hyacinthhas caused many economic, social and environmentalproblems. It blocks waterways, affects water transportfor agriculture and tourism, covers lakes and rivers,lowers the dissolved oxygen in water bodies and
* Biological Control Institute, Chinese Academy ofAgricultural Sciences, 12 Zhongguancun Nandajie,Beijing, 100081, China. Email: [email protected]
30
reduces aquatic production. It also affects the irriga-tion of agricultural fields.
As an invasive alien species, water hyacinth hasposed a great threat to biodiversity by competing withnative plants for water, nutrients and space. At Caohai,Dianchi Lake, in Yunnan province, southwesternChina, the plant diversity has been greatly reduced inthe past 30 years because of the competition by waterhyacinth and the increased water pollution. Thenumber of water plant species at Caohai has fallenfrom 16 in 1960, 8 in 1970, 5 in 1980, to 3 in 1990 (Wu1993) . The plants absorption of heavy metals causes asecond water pollution problem after they die and sink.In rural areas, after control by harvesting, hugenumbers of plants were always heaped together alongthe banks of rivers and allowed to decay, which greatlyaffected environmental quality (Wu 1993).
Water hyacinth has also caused a series of problemsto local society in China. People have difficulty intheir daily lives as it covers their rivers, ponds andlakes. The health of local people is threatened as waterhyacinth provides a habitat for mosquitoes and flies. Itis said that water hyacinth even creates a public secu-rity issue: dense and high water hyacinth plantsprovide a nice place for criminals to hide.
Utilisation of Water Hyacinth
During the 1950s–1970s, water hyacinth was widelyused for animal food in China, as at that time, theeconomy in rural areas was very depressed and therewas great shortage of food for animals. It was also
used for fertiliser in a few areas. Some people eventried to make paper from water hyacinth plants. Butsince the end of 1980s, and the economy improved, thenumber of people seeking to use water hyacinth hasfallen. The sole use of water hyacinth now, and in onlya few places, is for feeding ducks. In some environ-mental institutes, water hyacinth is used as a test plantin the purification of polluted water.
Control Status
Manual removal has been employed in most areas inChina in the past 10 years. It is estimated that morethan 100 million RMB yuan (US$12m) was spent onartificial control of water hyacinth each year but thepractice was neither economic nor effective. Mechan-ical control is used in only a few places, but it cannotprovide long term control. In some areas, herbicidessuch as Roundup and paraquat were used, but they areprohibited in some places where the water is used bypeople and animals.
Biological Control
In China, biological control activities for water hya-cinth were initiated in early 1995, when the BiologicalControl Institute (BCI) introduced two weevils Neo-chetina eichhorniae and N. bruchi from the USA andArgentina, respectively. Upon the request of the localgovernment, host range tests for the two weevils wereconducted in Kunming, Yunnan Province in 1995.Forty-six plant species from 23 families representinglocal economic, ornamental, and ecologically impor-tant plants (Ding et al. 1998) were tested. As they haddone previously in the USA, Australia, India and othercountries, host range tests showed the weevils attackedand completed their life cycles only on water hyacinthand they were safe to other local plant species.
After host-specificity tests, weevils were firstreleased at four rivers in Wenzhou and Zhejiang prov-inces in September 1996. Some 1000 individuals of amixture of the two weevil species were released at eachriver. The weevils established at all the four riverswithin one year of release. At Lincun River, about 50%of water hyacinth plants were killed in late spring 1998,while a native grass, Paspalum spp., recovered andoccupied the space where water hyacinth grew. Sincethen, the density of water hyacinth has varied season-ally between 10 and 50% of coverage of the water sur-face. By means of water flow, the weevils spreadrapidly to water bodies up to 40 km from the release siteby the summer of 2000 (Ding et al., unpublished data).
Figure 1. The distribution (shaded areas) of waterhyacinth in China. Black areas are wheredamage is greatest.
31
In 1998, the weevils were distributed to Fuzhou,Fuqing and Fuan cities of Fujian Province where waterhyacinth was a great disaster. They established quickly,as it was warmer than in Zhejiang Province. Significantcontrol has been achieved at several release sites but nodetailed survey results are available yet.
In the early summer of 2000, a colony of the weevilswas introduced to Ningbo, another city of ZhejiangProvince, where they were released in one river. Thereis some concern about the ability of the weevils tooverwinter there. Located at around 30°N, Ningbo isin the north of Wenzhou and Fujian. In some yearswhen it is very cold in winter, the water hyacinthplants die. In the next year, plants may regrow fromseed. Hence, tests have been planned to see if theweevils can overwinter in Ningbo.
In 2001 the weevils will be introduced to GuandongProvince which is one of the areas in China most seri-ously affected by water hyacinth damage.
Besides the weevils, a bug, Eccritotarsus catarin-ensis, was introduced into China in the early spring2000 from the Plant Protection Research Institute(PPRI), South Africa but, for unknown reasons, hadnot established its population after 4 months. BCI willintroduce it again later. Pathogen surveys have alsobeen started in Fujian and Zhejiang provinces in May2000. Several promising isolates have been screened.
Integrated Control
In order to control the weed rapidly, an integratedcontrol system was developed from 1996 by BCI sci-entists. Several herbicides, e.g. Roundup (41% IPAsalt of glyphosate) and Caoganlin (10% salt of glypho-sate), were screened to supplement the activity of theweevils. Bioassay tests showed that Roundup andCaoganlin had almost no adverse effect on the adults,pupae, larvae and eggs of the weevils. The tests of inte-grating Roundup at different concentrations withweevils indicated that herbicides had to be used at alower concentration than normal, so as to not kill theplants too rapidly and not deprive the insects of foodand habitat. The details of those tests were reported inthe first IOBC water hyacinth workshop in Zimbabwein 1998 (Ding et al. 1999).
Prospect
Water hyacinth is still a big problem in South China(even a new disaster in some areas) although greatefforts have been made to control it in the past 10
years. As more and more attention from central andlocal governments is paid to improvement of the envi-ronment, control of water hyacinth is becoming one oftheir objectives. Biological control will be employedin more and more areas, but more effort still needs tobe made to make the public and government officialsaware of the important role that biological control canplay in the solution of the weed problem.
BCI research on the biological and integratedcontrol of water hyacinth will focus on the followingsubjects in the next few years by means of national andinternational collaborations:
• Study of the factors influencing the level of controlof water hyacinth achieved by weevils, includingthe nutrients in the water body, lower temperaturesin winter, natural enemies of the weevils,competition from other aquatic plants such asPaspalum species etc.
• Distribution of the weevils into more areas andintroduction of new insects from abroad. Anagreement between the South African and Chinesegovernments has been signed for collaboration onwater hyacinth over the next three years. BCI willobtain help from PPRI for the importation of newnatural enemies e.g. Eccritotarsus catarinensis.
• Conduct of field tests of integrated control on alarge scale. The results from the tests in 1996–1998on integrating herbicides with weevils will beverified and amplified on a large scale in the field inSouth China so as to modify the integratedmanagement system.
• Continuation of the survey of pathogens in SouthChina and introduction of promising fungi fromabroad. More effort will be put into pathogenstudies. In BCI a pathogen laboratory has been setup for the study of control of the weed by thismeans. China’s strong background on developingbiopesticides and bioherbicides in the past 40 yearsshould help the laboratory to make good progress inthe near future.
Acknowledgments
The China National Natural Scientific Fund Com-mittee (No. 39770502) funded the research reportedhere. Special thanks go to Dr Lu Qing Guang, Mr ChenZhiqun and Mr Fan Zhongnan for their research assist-ance in the water hyacinth project at BCI.
32
References
Anon., ed. 1954. Taxonomy catalogue for China’s plants:families and genera. Beijing, Science Press.
Diao, Z., ed. 1989. Aquatic weeds in China. Chongqing,Chongqing Press. (in Chinese)
Ding, J., Wang, R., Fan, Z., Chen, Z. and Fu, W. 1995. Thedistribution and importance of water hyacinth in China.Chinese Journal of Weeds, 9, 49–51. (in Chinese withEnglish abstract)
Ding, J., Wang, R., Chen, Z. and Fu, W. 1998. Study on thefeasibility of using Neochetina eichhorniae and N. bruchito control water hyacinth in Dianchi Lake, Yunnan
Province of South China. In: Cheng Dengfa, ed., Prospectof plant protection in 21st century. China Science andTechnology Press, 660–664. (in Chinese)
Ding, J., Wang, R., Fan, Z. and Fu, W. 1999. Towardsintegrated management of water hyacinth with insectsand herbicides in southern China, In: Hill, M.P., Julien,M.H. and Center, T.D., ed., Proceeding of the 1st IOBCWater Hyacinth Working Group, Harare, Zimbabwe,November 1998, 142–147.
Wu, K. 1993. On the ecological problems of Dianchi Lake,Yunnan Province. China Lakes (Reservoirs) Newsletter,1, 47–49. (in Chinese)
33
Biological Control Initiatives against Water Hyacinth in South Africa: Constraining Factors,
Success and New Courses of Action
M.P. Hill* and T. Olckers†
Abstract
The success of biological control initiatives undertaken against water hyacinth in South Africa has been variable,despite the establishment of six natural enemy species (five arthropods and one pathogen) between 1974 and1996. By contrast, successful biocontrol was achieved in a relatively short time frame (4 years) on Lake Victoriain Uganda and in Papua New Guinea, using only the two insect agents, Neochetina eichhorniae and N. bruchi.The variable results achieved in South Africa have so far been attributed to variable climatic conditions,eutrophication of the aquatic ecosystems and interference from integrated control operations. However,hydrological features, notably the size of the water body, and techniques for establishing agents, may also affectthe degree of biocontrol. It is believed that biocontrol is more successful in larger water bodies where wind andwave action increase the mortality of agent-stressed plants. These considerations have prompted several coursesof action in South Africa, notably: (i) mass-rearing and re-releases of agents that failed to establish at specificsites; (ii) evaluation of the impact of the combinations of agents already established; (iii) development ofmanagement strategies in which biocontrol can be appropriately integrated with existing control operations; and(iv) search for additional agents that are effective under more temperate conditions. The success of theseinitiatives will ultimately rely on the extent to which water authorities and policy-makers become educated about,and come to accept, the principles of biological control.
THE biological control program against water hya-cinth in South Africa was initiated in 1973 and resultedin the release of the weevil Neochetina eichhorniae in1974 (Cilliers 1991). Three agents have since beenreleased, including another weevil Neochetina bruchiin 1989, the moth Niphograpta albiguttalis (= Same-odes albiguttalis) in 1990 and the mirid Eccritotarsuscatarinensis in 1996 (Julien and Griffiths 1998; Hilland Cilliers 1999). In addition, two agents were inad-
vertently introduced: the pathogen Cercosporapiaropi, which was first recorded in 1987, and the miteOrthogalumna terebrantis, recorded in 1989 (Cilliers1991). Despite the high number of established agents(the highest of all the countries involved with this pro-gram), the success of these initiatives has been vari-able. Although several water hyacinth populations inSouth Africa have been significantly reduced throughbiological control, notably in the Eastern Cape Prov-ince (Hill and Cilliers 1999), the results overall do notcompare with those obtained in Papua New Guinea(Julien and Orapa 1999) and more recently on LakeVictoria in Uganda (Cock et al. 2000).
Hill and Cilliers (1999) discussed several factors thathave constrained the impact of the arthropod agents inSouth Africa. These include: (i) cold winters, which
* Weeds Division, ARC–Plant Protection ResearchInstitute, Private Bag X 134, Pretoria 0001, South Africa.Email: [email protected]
† Weeds Division, ARC–Plant Protection ResearchInstitute, Private Bag X 6006, Hilton 3245, South Africa.Email: [email protected]
34
vastly increase the time taken to control the weed; (ii)highly eutrophic waters in which the weed thrives; (iii)periodic removal of the weed and natural enemy popu-lations through flooding and drought; and (iv) interfer-ence from other control methods, notably herbicideapplications. This situation has prompted severalcourses of action in South Africa, which include thesearch for additional natural enemies that are effectivein cooler areas and the development of managementstrategies in which biocontrol can be appropriatelyintegrated with existing control operations.
In this paper, we further discuss the above con-straining factors and suggest two additional factors(viz. the size of the water body and techniques forestablishing agents), which might also affect the bio-control of water hyacinth in South Africa.
Factors Affecting the Efficacy of Biocontrol
Variable climatic conditions
Water hyacinth populations are subject to a widerange of climatic conditions in South Africa,including: (i) high altitudes (above 1500 m), temperatesummer rainfall areas, where frosting occurs fre-quently during the colder months (May to August); (ii)coastal, Mediterranean winter rainfall areas, wherefrost is absent; and (iii) coastal, subtropical summerrainfall areas. Although all five arthropod agents havebecome established on water hyacinth throughout thisclimatic range, this is little doubt that the varying con-ditions affect their impact.
In the high elevation areas of South Africa (high-veld), the plants and insects remain dormant for up to5 months of the year (May–September). Despite this,there is some evidence that agent-induced stressinflicted on the plants during summer increases themortality of the plants which suffer cold stress duringthe following winter (Cilliers and Hill 1996). How-ever, plant populations increase rapidly with the onsetof spring (late September and October) while theresurging insect populations, which have to regeneratefrom considerably lower numbers because of cold-induced mortality and low reproductive output, do notreach damaging levels until the end of summer (Marchand April), only to ‘crash’ during the following winter.Consequently, unlike the situation in tropical and sub-tropical areas, the agents persisting in temperate areasseldom reach the population densities required toseverely stress the weed, and successful biocontrol
thus takes considerably longer. Unfortunately, waterauthorities often regard such time lags as unacceptableand the water hyacinth mats are invariably subjected toother control methods, notably herbicide applicationsand mechanical removal, which further reduces thenatural enemy populations (Center et al. 1999).
The Mediterranean climate typical of the WesternCape Province may also have had a negative impact onthe agents, which appear to have been ineffective inthis region. However, the reasons for this are unclear,as the effect of cool, wet and frost-free winters and hot,dry summers on the agent populations has not beendetermined. In addition, other factors such as floodingand eutrophication (see below) also limit the efficacyof biocontrol in this region.
By contrast, biocontrol has been considerably moresuccessful in the coastal and subtropical areas of theEastern Cape (EC) and Kwa Zulu-Natal (KZN) prov-inces. This has occurred in both integrated control pro-grams, such as at Lake Nsezi on the Nseleni River nearRichards Bay (KZN) (Jones and Cilliers 1999), andpure biocontrol programs, such as at New Year’s Damnear Alicedale (EC) (Hill and Cilliers 1999). How-ever, even in the subtropical areas of South Africa,eutrophication (see below) has hampered the efficacyof biocontrol.
Although not quantified, the range of climatic con-ditions under which water hyacinth occurs in SouthAfrica certainly has an effect on the natural enemypopulations and thus the degree of biocontrol.Whereas successful biocontrol usually takes 3–5 yearsin tropical areas (Harley 1990), it takes considerablylonger (8–10 years) under more temperate situations.As a remedy, insect species that have short generationtimes and which are capable of rapid populationincreases during the 6-month growing season of waterhyacinth in South Africa should be targeted forrelease. Such agents have already been identified inSouth America and include a petiole-mining dolico-podid flies Thrypticus spp., a delphacid Megamelus sp.and a dictyopharid Taosa sp.
Eutrophication of aquatic ecosystems
Many of the rivers and dams in South Africa receiverun-off which is highly polluted with nitrates and phos-phates arising from agricultural activities. Theseeutrophic waters enhance the growth of water hyacinthand other aquatic plant species, both native and intro-duced, to such a degree that aquatic weed problemsshould be regarded as a symptom of eutrophication. Apositive implication for biocontrol is that natural enemy
35
populations may proliferate because of higher qualityhost plants (Room 1990). Alternatively, the impact ofthe natural enemies may be negated by the extraordi-nary plant growth caused by rapid leaf production. Thisappears to be the case at Hammarsdale Dam (KZN)where both N. eichhorniae, established since 1989, andE. catarinensis, established since 1998, have reachedvery high population densities but appear to have hadlittle impact on the weed population.
Although Hammarsdale Dam occurs in a warm-temperate area where the insects are not affected byfrost, this seems to be negated by severe pollution.Indeed, during the summer months some 50% of thedam’s inflow is made up of effluent from textileindustries and a wastewater treatment plant and thisincreases to 100% of the inflow during winter.Eutrophic conditions ideal for water hyacinth popu-lations to proliferate thus persist throughout the year.Current post-release evaluations at this site have indi-cated that, although the density of the weed popula-tion has not been reduced, the two agents appear tohave reduced the size of individual plants. Otherfactors may thus have played a role at HammarsdaleDam. One explanation is that the system may be toosmall for wind and wave action to continually disturbthe weed mat and thereby enhance plant mortality(see below).
Four other agents have been released at Hammars-dale Dam— N. bruchi and C. piaropi in 1989 and N.albiguttalis and O. terebrantis in 1991—but none havebecome established. Possible reasons for this includeinadequate release techniques (see below) and host-plant incompatibility in the case of N. albiguttaliswhich is poorly suited to the tall plants with elongatedpetioles typical of this site. Further releases of thesespecies are under way.
The different agents established on water hyacinthhave differing plant requirements. Niphograpta albi-guttalis requires plants with actively growing, youngtissue and is therefore unlikely to establish on plantsgrowing under oligotrophic (i.e. unpolluted or unen-riched) conditions. Heard and Winterton (2000)showed also that N. bruchi is more damaging than N.eichhorniae under eutrophic conditions. In addition,Jamil and Hussain (1993) showed that uptake of heavymetals by the two Neochetina species reduced femalefecundity and might thus prevent their establishmentin weed populations that have assimilated high con-centrations of heavy metal pollutants. These consider-ations emphasise the importance of host plant qualitywhen trying to establish agents on water hyacinth.
A strategy for the biocontrol of water hyacinth atHammarsdale Dam should involve new approaches.These would include: (i) reducing the effluent inflowinto the dam; (ii) releasing large numbers of the better-suited N. bruchi to ensure establishment; (iii) allowingsufficient time for biocontrol to be effective; and (iv)manipulating the water level in the dam to allow peri-odic flushing of the system.
Interference from herbicide control operations
In South Africa, the control of water hyacinth reliesheavily on the application of herbicides, and this policyhas been antagonistic to biological control for two rea-sons. Firstly, certain herbicide formulations used on theweed in South Africa, especially those with high sur-factant content, cause high mortality in the natural ene-mies. Although N. eichhorniae was resistant to mostherbicide applications, those that contained diquat asan active ingredient were toxic to the weevil (Uecker-mann and Hill 2000). These authors also found that allherbicides tested, with the exception of one glyphosate-based product that contained no surfactants, were toxicto the mirid E. catarinensis. Secondly, herbicidaldestruction of water hyacinth populations, especially inimpounded systems, causes extensive mortality of thesessile immature stages and dispersal of the adultstages, when the weed mats start to sink. Re-infestationof these treated sites occurs via seed germination andisolated plants that were left unsprayed and the waterhyacinth populations proliferate in the absence ofnatural enemies (Center et al. 1999).
Solutions to these problems, currently under inves-tigation, include: (i) using herbicide formulations thatare less toxic to the natural enemies; (ii) re-inoculatingplants that are overlooked during herbicidal applica-tions; and (iii) accepting the concept of leavinguntreated ‘reserves’ to act as refugia for the agents.Ultimately, successful integrated control of water hya-cinth in South Africa will rely on a change in the atti-tude of water authorities. This will entail theiracceptance that the control of water hyacinth dependson reducing the level of nutrients flowing into thewater bodies, allowing sufficient time for biocontrol totake affect and limiting the use of herbicides, particu-larly formulations that are damaging to the agents.
Hydrological features
The influence of hydrological features on water hya-cinth infestations and subsequent biological control has
36
often been underestimated. This is illustrated by threerecent examples of successful biocontrol of water hya-cinth, namely the lagoons of the Sepik River in PapuaNew Guinea (Julien and Orapa 1999), Lake Kyoga inUganda (Ogwang and Molo 1999) and Lake Victoria inUganda (Cock et al. 2000). All three systems compriselarge, deep-water bodies with a wind fetch greater than2 km (Clayton 2000). In these situations, the twoweevil species reduce the size of the plants, the plantssit lower in the water and the weed mats loosen andfragment more easily. The mats are then further frag-mented by wind and wave action, which also killsmany plants and causes the mats to sink, as occurred atLake Victoria (Ogwang, pers. comm.). Alternatively,the small mats may be flushed out of the system, asoccurred down the Nile River off Lake Kyoga(Ogwang and Molo 1999) and down the Sepik River inPapua New Guinea (Julien, pers comm.).
In South Africa, many of the impoundments aresmall (< 100 ha), shallow (<10 m) basins and are there-fore not subject to wind and wave action. Although theagents can inflict severe damage on the plants, with upto 30 adult weevils per plant in some areas, the lack ofphysical stress on the mats prevents them frombreaking up and the plants from sinking. Furthermore,some areas in certain impoundments are too shallow(<0.3 m) for the plants to sink and the roots merely reston the substrate, as occurred at New Year’s Dam nearAlicedale. Lack of wind and wave action, coupledwith an inability to flush these impounded systems,has prevented the spectacular success observed inPapua New Guinea, Lake Victoria and Lake Kyogafrom being repeated in South Africa.
South African river systems that are infested withwater hyacinth but which have not been impounded,present a different problem for biological control. MostAfrican rivers are prone to periodic flooding anddrought, which cause unscheduled, sporadic removalsof both weed and agent populations. This results inwater hyacinth resurging from dormant seed banks and,in the absence of the agents, proliferating to reach pre-biocontrol levels. In these situations, redistribution ofthe natural enemies and close monitoring of the weedpopulations is necessary to restore biological control.
Techniques for establishing agents
The use of appropriate release techniques mayprove critical in ensuring the establishment of naturalenemies on water hyacinth. Establishment relies on therelease of large, healthy populations of the agents ontohealthy plants in the field. In South Africa, the release
of the two weevil species as adults has mostly ensuredestablishment, while the pathogen, mirid, mite andmoth are more likely to establish when individualplants, heavily infested with them, are placed into theweed populations. All releases must be made in shel-tered areas that are protected from disturbance by bothbiotic or abiotic factors. Numbers released have alsoproved crucial, since large or multiple releases have ahigher chance of ensuring establishment. Indeed, thevery low numbers (less than 100) of N. bruchi releasedat several sites in South Africa may well explain itsfailure to establish in some areas and its poor distribu-tion. Furthermore, the very large releases of Neo-chetina species carried out on Lake Victoria (greaterthan 100,000) appears also to have contributed to thespectacular success of biocontrol.
A series of dossiers on the rearing, release and mon-itoring of natural enemies for water hyacinth is beingproduced by CSIRO Australia. One has already beencompleted for the two Neochetina species (Julien et al.1999), while others are either in press (e.g. that on themoths, N. albiguttalis and Xubida infusella) or in prep-aration (e.g. that on the mite, O. terebrantis, and themirid, E. catarinensis). These publications willprovide essential information on the techniquesneeded to ensure the successful establishment ofagents for water hyacinth.
Successful Biocontrol: a South African Case History
One of the best examples of successful biocontrol ofwater hyacinth in South Africa is New Year’s Dam, a150 ha impoundment near Alicedale (EC). In 1990,when the weed mat covered some 80% of the dam,around 200 adult N. eichhorniae were released. Theweevils became established, spread throughout thepopulation and by 1994 had reduced the weed matcover to less than 10% of the dam’s surface area. Theremaining plants were small (10 to 20 cm tall) andunable to sink because of very shallow water.Niphograpta albiguttalis, O. terebrantis and E. cat-arinensis were released in 1996 but failed to establish,possibly because of both incorrect timing of therelease (middle of winter) and the very poor conditionof the surviving plants. By 1998, the weed mat coverhad increased to 80% of the water surface, but, with nofurther releases, N. eichhorniae once again reducedthis to around 10% by 2000. This system is thus con-sidered to be under biological control and three factorsappear to have contributed to this.
37
Firstly, the system is oligotrophic in that the sus-taining catchment is fairly small, sparsely populatedand does not support intensive agriculture or industry.Run-off into the dam is thus low in nitrates and phos-phates and even before the introduction of the weevils,the plants were small (<35 cm) and nutrient-stressed.The weed’s resurgence in 1998 may have been initi-ated by above-average rainfall in this semi-arid area,which significantly increased the nutrient input to thedam. A small resident weevil population, caused bythe reduced weed mat and poor quality of the plants,allowed the resurging mat to temporarily ‘escape’ theweevils, which then took 2 years to respond andrestore biocontrol.
Secondly, climate appears to have played a signifi-cant role in the success of the weevils. New Year’sDam is situated in a warm temperate region character-ised by spring and autumn rainfall, summer tempera-tures of 20–35°C and winter temperatures that seldomdrop below 10°C. Consequently, the life cycle of N.eichhorniae might be protracted during the wintermonths but their populations are not hit by frosts.
Thirdly, and most importantly, no other controlmethods have been employed at this site. The town ofAlicedale, which obtains all its water from the dam,supports a small community, and the weed has neverseverely affected the quality or quantity of water. Inaddition, the infestation does not threaten any infra-structure and is not regarded as a source of infestationfor other nearby catchments and rivers. Consequently,the national water authorities are under no pressure tocontrol the infestation in the short-term and are thusprepared to allow biocontrol to operate in isolation.
Discussion
Problems with biological control of water hyacinth arepresumably not unique to South Africa and are likelyto be experienced elsewhere in the world. Althoughthe biocontrol program in South Africa has been lesssuccessful than those implemented in other tropicalareas of the world, it has, nevertheless, lessened theoverall impact of water hyacinth. Besides the few sit-uations where water hyacinth infestations have beensignificantly reduced, the plants have generallybecome smaller in size. Indeed, some 20 years agoplants of 1 m and taller were frequently recorded whiletoday plants in mature stands seldom exceed 0.6 m onaverage (C. Cilliers, unpublished data). Smaller plantscause less-extensive mats, which pose less of a threatto infrastructure. In addition, the natural enemiesreduce the rate of mat expansion after disturbances,
notably flooding, manual removal and herbicide appli-cations. As a result, water authorities are able to reducethe number of herbicide applications at many of thecontrol sites, leading to considerable economic andecological savings.
The success achieved at New Year’s Dam has notbeen repeated elsewhere in South Africa and this hasprompted several courses of action. Firstly, there areseveral sites where some of the agent species havefailed to establish and these have been targeted forredistribution. Mass-rearing and re-releases are aimedat establishing the full suite of natural enemies at allsites throughout the country, to ensure that inappro-priate release methods used previously were not thecause of non-establishment. Secondly, the impact ofcertain agents on the weed, notably E. catarinensis, O.terebrantis and C. piaropi, is unknown. Laboratoryand field studies have been initiated to quantify theefficacy of these agents, both in isolation and in com-bination with the other species, and thereby facilitatethe development of improved management strategiesfor water hyacinth. Thirdly, additional agents areunder investigation, and recent surveys in northernArgentina (Cordo 1999) and the upper Amazonianregion of Peru (H. Cordo et al., unpublished data)have revealed several species that might be suitablefor release in South Africa. These include: (i) thegrasshopper Cornops aquaticum which is very dam-aging but not suitably host specific (Oberholzer andHill 2001), (ii) several species of the petiole-miningfly Thrypticus; (iii) the delphacid Megamelus; and (iv)the dictyopharid Taosa. These species have favour-able attributes, notably the fact that, despite their trop-ical origin, they thrive in the cooler regions ofArgentina (Buenos Aires province) suggesting adap-tations to more temperate climates. In addition, allhave short generation times (less than 40 days) andare thus capable of rapid population increases. Thesespecies may thus be suitable for release in the coolerareas of South Africa where rapid populationincreases during the summer months could causemore damage to water hyacinth populations than theagents currently established.
Although some 26 years have elapsed since thefirst release of a biological control agent againstwater hyacinth in South Africa, the program hasremained very active in researching additional waysof controlling the weed. However, the emphasis hasshifted from a purely biological to a more integratedmanagement approach, which includes aspects ofbiocontrol, herbicide applications, manual removal,hydrological control and nutrient control. The
38
success of this program will ultimately rely on theextent to which water authorities and policy-makersbecome educated about, and come to accept, the prin-ciples of biological control.
Acknowledgments
We thank Jan-Robert Baars, Hester Sparks andHelmuth Zimmermann (all from the Plant ProtectionResearch Institute) and Marcus Byrne (University ofthe Witwatersrand) for comments on the manuscript.The technical assistance of Hardi Oberholzer is grate-fully acknowledged. The water hyacinth program issupported by the ‘Working for Water’ program, an ini-tiative of the Department of Water Affairs and Forestryof South Africa, and the Water Research Commission.
References
Center, T.D., Dray, F.A., Jubinsky, G.P. and Grodowitz,M.J. 1999. Biological control of water hyacinth underconditions of maintenance control: can herbicides andinsects be integrated? Environmental Management, 23,241–256.
Cilliers, C.J. 1991. Biological control of water hyacinth,Eichhornia crassipes (Pontederiaceae), in South Africa.Agriculture, Ecosystems and Environment, 37, 207–217.
Cilliers, C.J. and Hill, M.P. 1996. Mortality of Eichhorniacrassipes (water hyacinth) in winter following summerstress by biological control agents. In: Moran, V.C. andHoffmann, J.H., ed., Proceedings of the IX InternationalSymposium on Biological Control of Weeds, 21–26January 1996, Stellenbosch, University of Cape Town,Cape Town, South Africa. p. 507 (abstract).
Cock, M, Day, R., Herren, H., Hill, M.P., Julien, M.H.,Neuenschwander, P. and Ogwang, J. 2000. Harvesters getthat sinking feeling. Biocontrol News and Information,21, 1–8.
Cordo, H.A. 1999. New agents for biological control ofwater hyacinth. In: Hill, M.P., Julien, M.H. and CenterT.D., ed., Proceedings of the 1st IOBC Global WorkingGroup Meeting for the Biological and Integrated Controlof Water Hyacinth, 16–19 November 1998, Harare,Zimbabwe, 68–74.
Harley, K.L.S. 1990. The role of biological control in themanagement of water hyacinth, Eichhornia crassipes.Biological News and Information, 11, 11–22
Heard, T.A. and Winterton, S.L. 2000. Interactions betweennutrient status and weevil herbivory in the biologicalcontrol of water hyacinth. Journal of Applied Ecology,37, 117–127.
Hill, M.P. and Cilliers, C.J. 1999. A review of the arthropodnatural enemies, and factors that influence their efficacy,in the biological control of water hyacinth, Eichhorniacrassipes (Mart.) Solms-Laubach (Pontederiaceae), inSouth Africa. African Entomology, Memoir, 1, 103–112.
Jamil, K. and Hussain, S. 1993. Biochemical variations inovaries of water hyacinth weevils Neochetinaeichhorniae (Warner) exposed to heavy metals. IndianJournal of Experimental Biology, 31, 36–40.
Jones, R. and Cilliers, C.J. 1999. Integrated control of waterhyacinth on the Nseleni/Mposa rivers and Lake Nsezi inKwazulu-Natal, South Africa. In: Hill, M.P., Julien, M.H,and Center T.D., ed., Proceedings of the 1st IOBC GlobalWorking Group Meeting for the Biological andIntegrated Control of Water Hyacinth, 16–19 November1998, Harare, Zimbabwe, 160–167.
Julien, M.H. and Griffiths, M.W. 1998. Biological control ofweeds: a world catalogue of agents and their target weeds.Fourth Edition. CABI Publishing, Wallingford.
Julien, M.H., Griffiths, M.W. and Wright, A.D. 1999.Biological control of water hyacinth. The weevilsNeochetina bruchi and Neochetina eichhorniae:biologies, host ranges and rearing, releasing andmonitoring techniques for biological control ofEichhornia crassipes. Canberra, ACIAR Monograph No.60, 87p.
Julien, M.H. and Orapa, W. 1999. Structure andmanagement of a successful biological control project forwater hyacinth. In: Hill, M.P., Julien M.H. and Center,T.D., ed., Proceedings of the 1st IOBC Global WorkingGroup Meeting for the Biological and Integrated Controlof Water Hyacinth, 16–19 November 1998, Harare,Zimbabwe, 123–134.
Oberholzer, I.G. and Hill, M.P. 2001. How safe is thegrasshopper, Cornops aquaticum for release on waterhyacinth in South Africa? In: These proceedings.
Ogwang, J.A. and Molo, R. 1999. Impact studies onNeochetina bruchi and Neochetina eichhorniae in LakeKyoga, Uganda. In: Hill, M.P., Julien, M.H. and Center,T.D.. ed., Proceedings of the 1st IOBC Global WorkingGroup Meeting for the Biological and Integrated Controlof Water Hyacinth, 16–19 November 1998, Harare,Zimbabwe, 10–13.
Room, P.M. 1990. Ecology of a simple plant–herbivoresystem: biological control of Salvinia. Trends in Ecologyand Evolution, 5, 74–79.
Ueckermann, C. and Hill, M.P. 2000. Impact of herbicidesused in water hyacinth control on the natural enemiesreleased against the weed for biological control. Report tothe Water Research Commission No. 78. (in press)
39
Recent Efforts in Biological Control of Water Hyacinth in the Kagera River
Headwaters of Rwanda
T. M. Moorhouse*, P. Agaba* and T. J. McNabb†
Abstract
As part of regional water hyacinth management activities in the Lake Victoria Basin that also involve Kenya,Tanzania, Uganda and several international partners, Rwanda is currently implementing efforts to rear andrelease the two Neochetina weevil species as biological control agents through coordination of trainingactivities and training visits made to Uganda and Tanzania. Weevils for release in Rwanda have come fromstocks maintained in Uganda.
The implementation of the biological control program within the Kagera River system of Rwanda is expectedto further support the long-term control of water hyacinth in the Lake Victoria Basin by reducing waterhyacinth biomass in source waters. Funding and technical support for the implementation of the biologicalcontrol program for water hyacinth in Rwanda are being provided by Clean Lakes, Inc. through a two-yearcooperative agreement with the United States Agency for International Development Greater Horn of AfricaInitiative through the Regional Lake Victoria Water Hyacinth Management Program.
APPROXIMATELY 13 years ago water hyacinth wasofficially recognised as having invaded the world’ssecond largest lake, East Africa’s Lake Victoria.During the ensuing years various management activi-ties have been implemented by Kenya, Tanzania andUganda with support from several international part-ners and donor organisations. Recently, these coun-tries and Rwanda have begun to coordinatemanagement efforts through regional organisationssuch as East African Cooperation (EAC), the LakeVictoria Fisheries Organization, the Lake VictoriaEnvironment Management Program, or through bilat-eral memoranda of understanding. Biological control
efforts using the weevils Neochetina eichhorniae andN. bruchi began in late 1995 through release effortsinitiated by Uganda that continue to date. Rwanda iscurrently implementing water hyacinth controlthrough rearing and release efforts assisted by CleanLakes, Inc. (CLI), under cooperative agreementfunding from the United States Agency for Interna-tional Development (USAID), and through coordina-tion of training activities and visits carried out inUganda and Tanzania. Weevil stocks maintained inUganda are the source of weevils imported intoRwanda.
The Lake Victoria Basin water hyacinth infestationextends to its uppermost point within the Kagera Riversystem to the headwaters of Mukungwa River tribu-tary, located approximately 50 km northwest ofKigali, Rwanda (F. Orach Meza, pers. comm.). TheMukungwa River is joined by the Nyaborongo River,keeping the latter’s name, until it merges with the
* Clean Lakes, Inc. – Uganda, Nile International ConferenceCenter, Room 235, PO Box 7057, Kampala, Uganda.Email: [email protected]
† Clean Lakes, Inc., P.O. Box 3186, Martinez, CA 94553,USA. Email: [email protected], or [email protected].
40
Burundi’s Ruvubu River system near Lake Rweru,along the Burundi border, to form the Akagera River,also known as the Kagera River (see Figure 1). Theentire Mukungwa/Nyabarongo/Kagera river system toLake Victoria is infested with water hyacinth, a lengthof over 500 km. Water hyacinth ultimately enters LakeVictoria in the form of mats torn away from the shore-line or as individual plants. There is at least one set ofmajor waterfalls along the Rwanda/Burundi border atRusoma, Rwanda, and a large swamp/lake complexalong the Rwanda/Tanzania border of the AkageraRiver where water hyacinth becomes damaged or iscaught in the swamp matrix, thus potentially reducingamounts travelling downstream. Downstream of thislarge swamp/lake system, which forms a large part ofthe Akagera National Park, the Akagera River changesdirection to an easterly course, becomes shared byTanzania and Uganda, and experiences a series of ele-vational drops near Kikagati, Uganda, where waterhyacinth again becomes damaged by turbulent waters.
Below Kikagati at a point approximately 160 kmfrom Lake Victoria, the river flattens and passes pri-marily through Tanzania where water hyacinth flour-
ishes along river banks, growing toward the rivercentre to a width of about 2 m from the shoreline.Water currents and velocity prevent water hyacinthfrom growing much beyond that with the exception insome bends, inlets or sloughs, or during periods ofdrought or flood. Considering that all rivers have twobanks, these 160 km of river therefore produce 320 kmof linear shoreline growth potential for the weed to awidth of approximately 2m, or a total of about 64 ha. Ithas been visually estimated by CLI staff that within 1km of Lake Victoria the daily rate of weed flowingdown the Kagera River ranges between 0.2 and morethan 1.5 ha/day (average 0.75 ha/day or 300 ha/year),depending on seasonal river flow. Others have esti-mated weed flow rates at 3.5 ha/week or 0.5 ha/day(Twongo and Balirwa 1995). If a growth rate model of1% per day were assumed, then these 64 ha growingalong the shoreline would generate about 0.64 ha ofnew weed growth/day. This is equivalent, on average,to the estimated daily inflows documented by CLIstaff in 1997 (unpublished data) and by Twongo andBalirwa (1995).
Kigati, Uganda
MukungwaRiver Bridge
L. Mihindi
NyabarongoRiver Bridge
L. Mpanga
Gashora Bridge
Rusuma
ISAR/KaramaWeevil RearingStation
L. Kivu
L. CyohohaSouth
L. Rweru
L. Sake
L. Mugesera
L. Muhazi
L. Rwanyakizinga
L. Thema
L. Bunyoni
I. Idjwi
Bugarama
Kibuye
Nyabisindu
RuhengeriGabiro
Kiziguro
Kayonza
Kibungo
Kagitumba
Kigali
Kagit
umba
R.
Aka
gera
R.
Ruvuv u R.
Ru
tshuru R .
Rusizi R.Nyabar ongo R.
BURUNDI
DEMOCRATIC REPUBLICOF THE CONGO
UGANDA
RWANDA
TANZANIA
0 20 kilometres
29°E
2°S 2°S
30°E 31°E
Kagera R.
Figure 1. Rivers and lakes of Burundi and adjacent areas of the Lake Victoria Basin
41
The Rwandan Biological Control Effort—a Summary
It was recognised that, in order to bring the water hya-cinth under management within the entire Lake Vic-toria Basin, a cooperative effort should be encouragedbetween the countries concerned. Recommendationswere made in various regional and EAC fora to includeRwanda and Burundi in efforts to manage water hya-cinth in the lake basin. In 1997, the governments ofRwanda and Uganda signed a memorandum of under-standing on common agriculture issues to cooperateon, among other things, water hyacinth management.Both governments committed their countries to fullcollaboration in management of the water hyacinthproblem.
During a Uganda Agricultural Policy Committeemeeting in July 1999, the Uganda National Agricul-ture Research Organization (NARO) sought supportfor Rwandans to be trained for implementing biolog-ical control of water hyacinth in the upper KageraRiver watershed.
In August 1999, EAC adopted a regional strategydocument on the control of water hyacinth and otherinvasive weeds in East Africa. This document coveredthe issue of water hyacinth control in the upper KageraRiver.
In September 1999, CLI staff visited Rwanda tohold discussions with the Director General, Institut desScience Agronomique du Rwanda (ISAR), to exploreinterest in training and release programs for biologicalcontrol of water hyacinth in the upper Akagera Riverwatershed. While there, CLI staff visited the Nya-barongo River, a tributary of the Kagera River thatflows south of Kigali, and reviewed the weed infesta-tion.
During November 1999, the Secretariat of theEAC, in close collaboration with the United NationsEconomic Commission for Africa, Kigali SubRegional office, held a workshop in Entebbe on waterhyacinth in the Lake Victoria Basin. One of the rec-ommendations of the workshop was that a programfor the biological control of water hyacinth should beimplemented in the upper Kagera River.
Through a USAID cooperative agreement, CLIfacilitated the training of Rwandan and Burundiangovernment officials in November 1999. The trainingwas led by Dr James Ogwang, head of biologicalcontrol programs at NARO’s Namulonge Agricultureand Animal Production Research Institute (NAARI)and by staff of the Uganda Ministry of Agriculture,
Animal Industries and Fisheries/Water Hyacinth Unit(MAAIF/WHU).
In May 2000, MAAIF/WHU and CLI staff visitedRwanda, in cooperation with ISAR officials, to iden-tify locations suitable for establishing the initialweevil rearing centre. Karama Animal Husbandry andFisheries Unit, one of the ISAR branch locations inRwanda, was selected because of its relative closenessto the Nyabarongo/Akagera River system. This unit islocated in the southern part of the country, approxi-mately 70 km southeast of Kigali in the Commune ofGashora on the shores of Lake Kilimbi.
In July 2000, MAAIF/WHU, CLI and ISAR offi-cials erected two weevil rearing tanks at ISAR/Karama. The tanks were filled with water and waterhyacinth.
In mid August 2000, the Rwanda Ministry of Agri-culture, Livestock and Forestry issued an authorisationallowing the Director General, ISAR, to import waterhyacinth weevils.
During 19–22 September 2000, ISAR/Karamaweevil rearing facility and CLI staff travelled to theweevil rearing sites at Kyakairabwa (Bukoba) andKyaka (Kagera River), Tanzania, to review and furtherstrengthen weevil-rearing techniques and experiencethrough observation and discussions with weevil-rearing technicians. On their return to Kampala, theyproceeded to the village sometimes known as Goma(near Kasensero) to a point approximately 1 kmupstream from the Kagera River outlet on Lake Vic-toria to observe water hyacinth growth. Weevildamage was noted on plants. The weevils presentmight have come from releases carried out by theGoma weevil rearing site, or have migrated upriverfrom Lake Victoria after releases carried out in theBukoba, Tanzania area, or from upriver releases atKyaka (quarterly releases since June 1998) under thesupport of the Lake Victoria Environment Manage-ment Project, or at Mulongo, Tanzania throughsupport from the International Fund for AgricultureDevelopment under the Kagera Agriculture and Envi-ronmental Management Program (KAEMP) (quar-terly releases since 1999). The ISAR/CLI teamunderstood that KAEMP had also made two weevilreleases on the Tanzanian side of the Akagera River,opposite Rusoma, Rwanda, though no dates weregiven and details remain sketchy.
In September 2000, the Deputy Director – Researchof the Ugandan NARO, issued a letter grantingapproval for Rwanda to export weevils from Ugandaand gave permission for NAARI to collect and prepareweevils for transit to Rwanda.
42
On 25 September, weevils were collected in coop-eration with NARO from the NAARI weevil rearingtanks under the direction of the head of biologicalcontrol programs and assisted by NAARI, ISAR, andCLI staff. The numbers of weevils collected for trans-port were as shown in Table 1.
In order to help monitor the efficacy of weevilreleases, satellite images were scheduled for acquisi-tion in late September or early October 2000,depending on cloud cover. IKONOS 1-metre PANand 4-metre multispectral band data were to be col-lected in collaboration the United States GeologicalSurvey–EROS Data Center. One image each will beacquired for the small lakes, Lake Mpanga and LakeMahindi, in the Akagera River area of AkageraNational Park (eastern Rwanda) at the upstream anddownstream ends of the swamp/lake complex, respec-tively. While influences such as flooding, drought andpathogens may also lead to declines in water hyacinth,it is expected that these images will provide a baselinesample for tracking water hyacinth cover and distribu-tion before and after weevil release. On-the-groundobservations will be made along other sections of theriver.
On 27 September 2000, ISAR and CLI staff trav-elled by air from Entebbe, Uganda to Kigali, Rwanda,with the consignment of Neochetina spp. Upon arrivalin Kigali, they were met and transported to the ISARKarama weevil rearing facility in order to inoculate the
weevils into water hyacinth in previously establishedtanks. Approximately 800 weevils were placed in thewater hyacinth plants of the two tanks. The two weevilspecies were deliberately mixed when placing them inthe tanks. The Karama weevil rearing site is within 10km of the Gashora Bridge on the Nyabarongo River,which is expected to become one of several weevilrelease sites within Rwanda.
On 28 September, approximately 25 weevils of eachspecies were released in a small depression, LakeKiruhura, in the Nyabarongo River floodplain,approximately 2 km east of the Nyabarongo RiverBridge, 20 km south of Kigali. This seasonal lake, justover a hectare in area, lies approximately 200 m southof the Nyabarongo riverbank. It was about 60%covered with water hyacinth of estimated averageheight 30–40 cm. About 40% of the plants were inflower. As a result of drought in much of Rwanda atthe time, rivers, pools, and depressions were experi-encing very low water levels. It was expected, how-ever, that the rainy season would commence shortlyover the entire country, as rains had at release timebeen reported in the upper Mukungwa/Nyabarongorivers of northern Rwanda.
Conclusion
Studies and evaluations of the effectiveness of thebiological control program will continue, and updateswill be provided as data become available. The nextreport on the campaign is scheduled for release inJanuary 2001.
Reference
Twongo, T. and Balirwa, J.S. 1995. The water hyacinthproblem and the biological control option in the highlandlake region of the Upper Nile basin—Uganda’sexperience. Paper presented at the Nile 2000 conference:Comprehensive Water Resources Development of theNile Basin: Taking Off. 13–17 February 1995, Arusha,Tanzania.
Table 1. Numbers of water hyacinth weevils collectedin Uganda in September 2000 for release inRwanda
Chevroned water hyacinth weevil
(Neochetina bruchi)
Water hyacinth weevil (Neochetina
eichhorniae)
Females 117 330
Males 127 280
Subtotal 244 610
Total 854
43
Ongoing Activities in the Biological Control of Water Hyacinth in Egypt
Y.H. Fayad,* A.A. Ibrahim,* A.A. El-Zoghby* and F.F. Shalaby†
Abstract
As in many other tropical and subtropical countries, the aquatic floating weed water hyacinth causes seriousproblems to various types of water bodies in Egypt. The total infested area is estimated to be 487 km2 coveringmost of the drainage and irrigation canals in different governorates of Egypt, and about 151 km2 covering lakes.The total amount of water loss by evapotranspiration from water hyacinth infested areas was estimated to be3.5 billion m3 per year. This amount is sufficient to irrigate about a further 432 km2 every year.
During the period 1978 to 1984, two weevils, Neochetina eichhorniae Warner and N. bruchi Hustache, wereintroduced into Egypt and studied under quarantine conditions. Host-specificity tests proved the safety of bothweevils for release as biocontrol agents for water hyacinth. No authorisation for release was given until 1999,when a biological control program to be financed by the French Government was approved. In May 2000, 3004weevils—1118 N. eichhorniae and 1886 N. bruchi—were collected from water hyacinth infested sites in FortLauderdale, Florida, and transferred to Egypt for rearing and multiplication in an aquatic weed greenhouse.Field releases of both species in Egypt began in August 2000.
WATER hyacinth, Eichhornia crassippes (Mart.)Solms (Pontederiaceae), was first recorded in Egypt bySimpson (1932). During the period 1978–1984, incooperation with the United States Department ofAgriculture, Fort Lauderdale, Florida, two weevil spe-cies, Neochetina eichhorniae Warner and N. bruchiHustache, were introduced into Egypt and studiedunder quarantine conditions through a PL480 project.These beneficial weevils have been introduced andreleased in several countries in the world (Julien andGriffiths 1998), such as the USA (Perkins 1973; Center1982), Australia (Wright 1979) and Sudan (Besher andBennett 1985), to control water hyacinth. Host-specif-icity tests proved the safety of both weevils for releaseas biocontrol agents (Julien et al. 1999; DeLoach 1976;
Fayad 1982, 1999). No authorisation for release inEgypt was given until 1999 when the Egyptian Min-istry of Agriculture approved the introduction of thetwo Neochetina weevils for release on water hyacinth.A biological control program, financed by the FrenchGovernment, started in January 1999.
In May 2000, 3004 weevils—1118 N. eichhorniaeand 1886 N. bruchi—were collected from water hya-cinth near Fort Lauderdale, Florida, USA. They weretransferred to Egypt for rearing, multiplication andrelease in the northern lakes Mariout (AlexandriaGovernorate) and Edko (Beheira Governorate).
Methods
Construction of aquatic weed greenhouse and growing of water hyacinth
A light and temperature-controlled aquatic weedgreenhouse of dimensions 15 × 7 m, containing 9 cir-
* Department of Biological Control, Plant ProtectionResearch Institute, ARC, Dokki, Egypt. Email: [email protected]
† Plant Protection Department, Faculty of Agriculture atMoshtohor, Zagazig University, Egypt.
44
cular water pools of 160 cm diameter and 100 cmdepth each, were constructed in the Department ofBiological Control, ARC at Giza. Water supply anddrainage was provided to each pool. The greenhousewas quarantine secure and supplemented with adressing room, bathroom and laboratory. The poolswere filled with tap water to 80 cm depth and left for 3days to remove any undesired dissolved gases beforeintroducing water hyacinth plants collected from theRiver Nile or irrigation canals. Water hyacinth plantswere washed thoroughly with tap water before place-ment in the pools.
Fifteen grams of a soluble NPK fertilizer containingmicro nutrients (Polyfeed – Haifa Chemicals Ltd) plus70 g nitrogen and 10 g iron, were added to each poolmonthly or as required, as indicated by leaves turningyellow or a blue colour appearing on the roots. Thetemperature in the greenhouse was set at an average ofabout 27.8°C. Plants were washed daily using tapwater to prevent aphid and mite infestations. Waterwas added to the pools whenever needed.
Collecting Neochetina bruchi and N. eichhorniae for introduction into Egypt
A collecting trip was made to Fort Lauderdale,Florida, USA, during the period 14–19 May 2000.Individuals of both Neochetina bruchi andN. eichhorniae were collected by handpicking fromwater hyacinth plants either from the shore or a boat.Collected weevils were kept in plastic containers pro-vided with tissue paper and water hyacinth leaves. Thecontainers were transferred to the laboratory, weevilsseparated into species and stored in the refrigerator at10°C. Over 5 days, a total of 3004 weevils of both Neo-chetina species was collected. Samples from the col-lected weevils were taken, dissected and micro-scopically examined for insect-disease detection.Weevils were placed in screw-capped carton tubes forshipping, 200–300 weevils per tube. The tubes werestored in the laboratory at 18°C for 2 days then hand-carried to Egypt. Upon arrival, insects were inspectedby quarantine officers at Cairo Airport then transferredto the quarantine room attached to the aquatic weedgreenhouse and stored in a refrigerator at 10°C. Thefollowing day, the insects were examined and healthyweevils were placed on healthy water hyacinth plantsplaced in the 9 water pools in the greenhouse.
N. bruchi adults were released on water hyacinth in6 pools, while N. eichhorniae were released in theother 3 pools. Each pool was allocated 200–250 wee-
vils. The rest of the weevils were stored in the refrig-erator at 10°C.
The pools were examined daily for feeding spotsand symptoms of weevil activity.
Harvesting the weevils from water hyacinth plants in the greenhouse
The first generation started to emerge in the firstweek of August 2000. During the daytime, manyadults were found on the floor of the greenhouse,driven out of the pools by overpopulation. Theseadults were collected. Furthermore, a metal ring net of120 cm diameter was placed on the water hyacinthplants in the pool and pushed down to submerge them.Floating adults were then collected using a medium-size strainer. The weevils collected each day werecounted and placed in square, plastic containers fur-nished with tissue paper and provided with fresh greenwater hyacinth leaves. A window of about 8 × 4 cmcovered with wire gauze was made in the cover lid forventilation. The containers were stored in the refriger-ator at 10°C until release of the weevils.
Field release of N. bruchi and N. eichhorniae
Harvested weevils were transferred in an icebox tothe release sites at Mariout and Edko lakes. Both lakeswere chosen as they were heavily infested with waterhyacinth. Airboats, rubber boats and sometimesnarrow wooden fishing boats were used for transfer inthe lakes. Release details are given in Table 1.
Results
Establishment of the weevils in the greenhouse
Feeding scars and weevil activities includingmating and oviposition were observed 2 days afterrelease. The damage to water hyacinth plants becamemore obvious each day, and new plants were addedperiodically or whenever needed. Leaf and petiolesamples were taken, dissected and examined under thebinocular microscope. Weevil eggs could be seen.Three weeks later, root inspection indicated the pres-ence of different instar larvae attached to the plantroots. In late July, cocoons containing pupae were alsofound, attached to the submerged roots of the plants. Inthe first week of August, the first generation started toemerge, indicating success in rearing the introducedweevils under greenhouse conditions.
45
Release of N. bruchi and N. eichhorniae in Mariout and Edko lakes
Some 6573 adults of both Neochetina species werereleased in Mariout and Edko lakes (Table 1). Further-more, water hyacinth plants supporting different Neo-chetina life stages were distributed on water hyacinthinfestations at certain sites in the lakes.
Follow up of the establishment of the released weevils
Weekly visits to the release sites in each lake wereconducted to determine the establishment and spreadof the weevils. Many feeding scars were observed onwater hyacinth leaves, suggesting the establishmentof Neochetina weevils at the release sites in bothMariout and Edko lakes, but more time is needed toconfirm this.
Discussion and Conclusion
Despite the studies conducted in Egypt since 1978 onthe use of insects for the biological control of waterhyacinth with results indicated the safety of releasingboth N. bruchi and N. eichhorniae (Fayad 1982, 1999),no authorisation for release was given until 1999.
The success of rearing both weevils under green-house conditions and symptoms of feeding scars grad-ually increasing after releasing the weevils in bothlakes are very promising signs. By introducing andreleasing N. bruchi and N. eichhorniae, the authorsannounce that, in August 2000, Egypt joined othercountries in applying biological control to water hya-cinth. In combination with other control methods it is
hoped to gain an acceptable level of water hyacinthcontrol that keeps the population under the economicthreshold. Several other known biocontrol agents haveto be introduced. A search for, and study of, newagents needs to continue.
Acknowledgments
The authors wish to thank the French Government forfinancing the project entitled ‘The biological controlof water hyacinth in Egypt’, which covered all travelexpenses for collecting in the USA and participation inthis working group meeting. Great appreciation isextended to the scientists of USDA at Fort Lauderdale,Florida, for their sincere assistance provided duringthe collecting trip on 14–19 May 2000.
References
Besher, M.O. and Bennet, F.D. 1985. Biological control ofwater hyacinth on the White Nile, Sudan. Proceeding ofthe VI International Symposium on Biological Control ofWeeds, Vancouver, Canada, 1984, 491–496.
Center, T.D. 1982. The water hyacinth weevils: Neochetinaeichhorniae and N. bruchi. Aquatics, 4, 8, 16–19.
DeLoach, C.J. 1976 . Neochetina bruchi, a biologicalcontrol agent of water hyacinth; host specificity inArgentina. Annals of the Entomological Society ofAmerica, 69, 635–642.
Fayad, Y.H. 1982. Insects for the control of water hyacinthin Egypt. Ph.D. thesis, Plant Protection Department,Faculty of Agriculture, AL-Azhar University, 161p.
—1999. Water hyacinth infestation and control in Egypt.Proceedings of 1st IOBC Working Group Meeting for theBiological Control of Water Hyacinth 16–19 November1999, Harare, Zimbabwe, 106–110.
Table 1. Releases of N. bruchi and N. eichhorniae in Mariout and Edko lakes
Date of release Site of release No. of released adults Total No. of release sites
N. bruchi N. eichhorniae
15.8.2000 Edko Lake 1395 692 2087 3
21.8.2000 Mariout 668 332 1000 3
infested plants distributed 1
30.8.2000 Edko Lake 702 298 1000 3
infested plants distributed 1
11.9.2000 Mariout 812 490 1302 3
19.9.2000 Edko Lake 787 397 1184 3
infested plants distributed 2
46
Julien, M.H. and Griffiths, M.W. 1998. Biological control ofweeds. A world catalogue of agents and their target weeds,4th ed.Wallingford, UK, CAB International, 223p.
Julien, M.H., Griffiths, M.W. and Wright, A.D. 1999.Biological control of water hyacinth. The weevilsNeochetina bruchi and Neochetina eichhorniae:biologies, host ranges, and rearing, releasing andmonotoring techniques for biological control ofEichhornia crassipes. ACIAR Monograph No. 60, 87p.
Perkins, B.D. 1973. Potential for water hyacinthmanagement with biological agents. Proceeding of the
4th Conference on Ecological Animal Contributions byHabitat Management, Tallahassee, Florida, 1972, 53–64.
Simpson, N.D. 1932. A report on the weed flora of irrigationcanals in Egypt. Cairo, Egypt, Ministry of Puplic Worksand Water Resources, Government Press, 124p. (inArabic).
Wright, A.D. 1979. Preliminary report on damage toEichhornia crassipes by an introduced weevil at a centralQueensland liberation site. Proceedings of the ThirdAustralian Applied Entomology Research Conference,25–27.
47
Progress with Biological Control of Water Hyacinth in Malawi
P.M. Phiri*, R.K. Day†, S. Chimatiro‡, M.P. Hill#, M.J.W. Cock**, M.G. Hill†† and E. Nyando*
Abstract
Water hyacinth appeared in southern Malawi during the late 1960s, and spread slowly northwards in theLower Shire River, but in 1995 it was found in the Upper Shire River, just south of Lake Malawi. It is nowpresent in most parts of the Shire River, and in a number of other locations, including the far north of thecountry. Biological control was initiated in 1995 under a UK Department for International Development-funded project, and is now being continued through a World Bank-funded project. About 200,000 Neochetinahave been reared and released, mainly in the Shire River, but recently at other sites outside the Shire. Thebeetles are well established in the Shire, though establishment and subsequent population build-up has beenfaster in the Lower Shire than the Upper and Middle Shire. Water hyacinth infestation in the Shire River isnow less than it was two years ago, but it is too early to conclude that this is the result of the biological controlcampaign. As new infestations appear elsewhere in the country, biological control agents will be released tolimit build-up of the weed.
FISHERMEN in the southern tip of Malawi report thatwater hyacinth first appeared there in the Shire Riverin the 1960s, and suggest that it may have arrived fromacross the border during floods (Harley 1991; Chi-matiro and Mwale 1998), a reasonable hypothesisgiven that it had been present in Zimbabwe and theZambezi River for many years before its discovery inMalawi. It subsequently spread slowly northwards,and by 1980 was present at the southern end of Ele-phant Marsh (Blackmore et al. 1988) (near Makhanga;
see Figure 1). By 1991 it had reached the northern endof Elephant Marsh, south of a Chikwawa (Terry 1991).In 1995 it was discovered in the Upper Shire Rivernorth of Mangochi, although surveys indicated it wasnot present between Lake Malombe and Chikwawa atthat time (Hill et al. 1999), suggesting that it had beenaccidentally introduced to the Upper Shire.
The Shire River can be divided into four sections(Table 1), but only the Murchison Rapids section isunsuitable for the weed. As well as occurringthroughout the Shire, water hyacinth is now present ata number of locations across the length of Malawi,including Blantyre, Lilongwe River, Salima, Nkhota-kota, south of Nkhata Bay, and north of Karonga. Thereare unconfirmed reports from other locations, includingthe Songwe River along the northern border with Tan-zania, so the weed is clearly now widely distributed.However, in most places outside the Shire River infes-tations are generally relatively small.
* Fisheries Department, PO Box 47, Mangochi, Malawi.† CAB International Africa Regional Centre, PO Box 633,
Village Market, Nairobi, Kenya.‡ Fisheries Department, PO Box 593, Lilongwe, Malawi.# Plant Protection Research Institute, Private Bag X134,
Pretoria 0001, South Africa.** CABI Bioscience Switzerland Centre, Delemont,
Switzerland.†† CABI Bioscience UK Centre, Silwood Park, Ascot,
Berkshire, SL5 7TA, UK.
48
A project was commenced in late 1995, funded bythe UK Department for International Development(DFID), which focused on the Shire River. The projecthad four components: biological control, publicawareness and community participation, socioeco-nomic evaluation of the problem, and assessment ofthe weed’s environmental impact, and these weredescribed at a workshop in Harare, Zimbabwe in 1998(Hill et al. 1999). The project ran for a little over threeyears, after which there was a hiatus of nearly a yearbefore a new project started, funded by the World
Bank. The new project again has several components,but biological control remains the main thrust in thestrategy for long-term control. In this paper we providean update on progress with the biological control ofwater hyacinth in Malawi.
Methods
Rearing and release of biological control agents
Rearing tanks for Neochetina spp. were establishedat Fisheries Department facilities at Makhanga andrearing was started using N. bruchi and N. eichhorniaehand-carried from Zimbabwe in September 1996. Partof the initial importation was used to make small-scalereleases. Tanks were set up at Mangochi in May 1997using insects from the first harvest at Makhanga. Ini-tially, both units had 10 tanks, but later 5 were added atMakhanga, and 10 at Mangochi, though the tanks atMangochi have not been used for Neochetina rearingcontinuously.
Methods used were adapted from those used inSouth Africa. Rearing tanks were cylindrical, with adiameter of 265 cm and height of 67.5 cm, and so con-tained approximately 3000 litres of water. Each tankwas placed on a concrete plinth, with drainage chan-nels between tanks leading to a soakaway. Waterlevels were checked daily and topped up as required.Once a fortnight the water was replaced using waterpumped from the river. At replacement 500 g of ureaand 250 g fertiliser (NPK: 6-18-6 or other as available)were added to each tank. Dead leaves and plants wereremoved as necessary. Harvesting was undertakenabout once a month, and harvested beetles werecounted by species and sex then released.
Mangochi
L. Malombe
Matope
Chikwawa
Liwonde
Makhanga
Karonga
Lilongwe
Blantyre
100 km
Nkhata Bay
Nkhotakota
Salima
Table 1. Sections of the Shire River (adapted from Crossley 1980, quoted Blackmore et al. 1988)
Section Between Gradient mm/km
Features
Upper Shire Lake Malawi and Lake Malombe ~100 Fisheries
Middle Shire Lake Malombe and LiwondeLiwonde and Matope
1696
Liwonde National ParkBarrage at Liwonde
Murchison Rapids Matope and Chikwawa ~5000 Hydroelectric power stations
Lower Shire Chikwawa and southern border 250 Sucoma sugar plantationElephant and Ndinde marshes (major fisheries)
Figure 1. Map of Malawi showing places referred to inthe text
49
Eccritotarsus catarinensis and Niphograpta albi-guttalis were imported in May 1997, May 1998,January 1999 and April 2000, hand-carried fromSouth Africa. Initially, imported insects werereleased, but subsequently a part of each importationwas released and part used to set up cultures in tanksat Mangochi.
The mite Orthogalumna terebrantis was alreadypresent on the weed in the Lower Shire, having accom-panied the weed from the Zambezi where it wasreleased in Zambia in the 1970s (Julien and Griffiths1998). It was redistributed onto water hyacinth in theUpper Shire during 1996 and 1997 and is now wellestablished on the weed throughout its range in the ShireRiver. Mite-infested leaves are being redistributed onnew infestations of the weed as they are discovered.
Monitoring
Currently there are 14 sites at which the impact ofbiological control is being monitored, 3 in the UpperShire, 3 in the Middle Shire, 5 in the Lower Shire, 2 onLake Malawi, and 1 in Blantyre. Monitoring is under-taken once every 2 months, and on each occasion 30mature plants are selected at random and the parame-ters listed in Table 2 recorded for each plant.
Results and Discussion
Rearing
Figure 2 plots the Neochetina harvested at the tworearing units. By mid 2000 the Makhanga unit hadproduced over 100,000 and the Mangochi unit about90,000. Initial harvests at Makhanga were high as thetanks had been running for 8 months before har-vesting commenced, so populations had reached highlevels. During 1999, production was intermittent asthe DFID project had ended and the new project hadnot yet started.
Table 4 shows that, at both units, there has been aslight excess of females in both species. At Makhanga,where the climate is hotter, slightly more N. bruchihave been produced than N. eichhorniae, while atMangochi, almost two-thirds of production has beenof N. eichhorniae. Rearing of both Eccritotarsus andNiphograpta has been unsuccessful: after 1–2 genera-tions the populations in the rearing tanks have died outfor reasons that are not clear.
Releases
The first releases of Neochetina were made in Sep-tember 1996, from the first importation. The first
Table 2. Parameters recorded during impact monitoring
Parameter Description
Longest leaf The length of the longest petiole plus lamina on the plant
Root length The maximum length of the root system
Lamina length for leaf 2 Leaf 2 is the second youngest/2nd most recently opened leaf.
Lamina width for leaf 2 At widest part of lamina
No. of leaves Includes sick leaves but not leaves on any daughter plants
No. of ramets The number of daughter plants attached to the sampled plant
No. of beetles The numbers of adult weevils of the two species released
Leaf 2 scars The number of weevil feeding scars on leaf 2
Leaf 2 mites The mite damage score, using the system in Table 3
Leaf 2 pathogens Damage caused by pathogens on leaf 2, using the same scale as for mite damage (Table 3)
Leaf 4 mites, pathogens As for Leaf 2
Leaf 5 mites, pathogens As for Leaf 2
Other agents Presence/absence of Eccritotarsus and Niphograpta
50
releases of insects reared in Malawi were made in May1997, and continued regularly until early 1999. Har-vesting and thus releases during 1999 were intermit-tent, but with the start of the new project in late 1999,harvesting and releases have again continued regu-larly. Including the first releases, by mid 2000 over190,000 weevils had been released. Figure 3 shows theproportion of insects released in different areas.Within an area a number of different release sites havebeen used.
The rationale for the pattern of releases between thedifferent areas is as follows:• More releases have been made in the Upper Shire
River as it was expected that populations would becarried downstream to the Middle and Lower Shire.
• During 1999, the time between the two projects, allof the weevils rearing at the Mangochi site werereleased in the Upper Shire.
• Establishment and build-up of populations in theLower Shire have been faster than in the Upper andMiddle Shire, requiring fewer releases.
• The first project focused on the Shire River. Sincethe start of the current project, releases have beenmade in Salima, Nkhotakota, and Chiwembe dam,Blantyre. The last site is only about 4 ha, but haslarge healthy plants due to pollution of the inflow(Limbe River). Further releases are being made in the Middle Shire
where establishment has been slow. No furtherreleases are required elsewhere unless sites are discov-ered where the beetles are absent.
To date a little over 5000 Eccritotarsus have beenreleased at sites in the Upper, Middle and Lower Shireand at Blantyre. About 800 Niphograpta have beenreleased in the Upper and Middle Shire.
Jan
98
July
98
Jan
99
July
99
Jan
00
July
00
0
2000
4000
6000
8000
10000
12000
14000
16000
No
. wee
vils
N. bruchiN. eichhorniaeTotal
Date
Mar
97
Sep
97
Mar
98
Sep
98
Mar
99
Sep
99
Mar
00
0
2000
4000
6000
8000
10000
12000
14000
16000
No
. wee
vils
N. bruchiN. eichhorniaeTotal
Date
(a) (b)
Figure 2. Neochetina harvesting at the two production units: (a) Mangochi; (b) Makhanga
Table 4. Percentage of Neochetina spp. production by sex and species at the two rearing units
Unit N. bruchi N. eichhorniae
Male Female Total Male Female Total
Makhanga 46.9 53.1 53.7 47.7 52.3 46.3
Mangochi 45.8 54.2 36.3 47.6 52.4 63.7
Table 3. Mite and pathogen damage scores
Score % of leaf occupied/damaged
0 0%
1 <5%
2 6–25%
3 26–50%
4 51–75%
5 75–100%
51
Monitoring
Here we present an example of the data-sets beingcollected. Figure 4 shows the mean number of weevilfeeding scars on leaf 2 for one site each in the Lower,Middle and Upper Shire River.
The beetles are established at all monitoring sites.Damage by the weevils has increased in the UpperShire in the last two years. In Lake Malombe and theMiddle Shire, damage levels have remained low,while in the Lower Shire there has been a build-up tohigher levels than in the Upper Shire. In recent monthsa reduction in weevil damage has been seen at someLower Shire sites, and this appears to be associatedwith the reduction in plant height, though it may be asimultaneous response to an environmental variablerather than a causal link.
From 1998 to 2000 there has been some decrease inplant height, but this is not matched by a reduction in
laminar area for leaf 2 or leaf number. If biologicalcontrol is working, we would expect a general reduc-tion in plant vigour and thus size.
Mite damage was generally higher in 1998 than sub-sequently, though there has been some increase in2000. Interestingly, the same pattern has occurred fordamage by pathogens, and this suggests that the mitesmay be facilitating infection by pathogens. However,both control agents may be responding to the sameenvironmental conditions. No evidence has beenfound for establishment of either Eccritotarsus orNiphograpta.
At all the sites on the Shire River there appears tohave been a reduction in the infestation of water hya-cinth, and at one site in the Lower Shire monitoring hasceased as there is now so little water hyacinth presentthat monitoring is impractical. In the Upper Shire thereduction of the weed appears to have coincided withan increase in cover by the sedge Pycreus mundtii, andRother and Twongo (1999) have suggested that thewater hyacinth is stimulating a succession in which itis being replaced by Pycreus and Ludwigia.
Conclusion
Neochetina spp. are well established in most parts ofthe Shire River, and numbers at some sites have built-up to levels at which a significant impact can beexpected on water hyacinth infestations. Hill et al.(1999) suggested that impact might become visible by2000–2001, and certainly the population of water hya-cinth in the Shire River is less than it was two yearsago. Fishermen in the Lower Shire are crediting theweevils with this reduction, but while this is pleasing,more data need to be collected to confirm this view.
In other parts of Malawi, new infestations can beexpected to occur. In some cases it may be possible forlocal communities to effect control by manualremoval—there are already some cases of thisreported. At the same time, the long-term strategyremains centred on biological control, and Neochetinaspp. and Orthogalumna will be released on significantnew infestations as they are reported. It is hoped thatEccritotarsus and Niphograpta will also becomeestablished.
As Julien and Orapa (1999) concluded, a successfulbiological program requires expertise, appropriatetraining and capacity building, staff and resources overan adequate period. We are confident that these ingre-dients are all present in the Malawi project, and so weare optimistic that the program will be a success.
0
5
10
15
20
25
30
Apr 98 Oct 98 Apr 99 Oct 99 Apr 00
Sample date
Wee
vil f
eed
ing
sca
rs, l
eaf 2
Upper Shire
Middle Shire
Lower Shire
Figure 4. Weevil feeding scars at three monitoringsites in the Shire River.
Figure 3. Releases of Neochetina eichhorniae andNeochetina bruchi in the Shire River.
Salima1%
Nkhotakota3%
Lake Malombe3%
Lower Shire14%
Upper Shire47%
Middle Shire17%
Blantyre15%
52
Acknowledgments
The first part of this work was funded by the UKDepartment for International Development throughthe Crop Protection Programme of its RenewableNatural Resources Knowledge Strategy. The currentproject is funded by the World Bank as part of theEnvironmental Management Programme. We thankthe Director of Fisheries and staff of the FisheriesDepartment for their support and assistance.
References
Blackmore, S., Dudley, C.O. and Osborne, P.L. 1988. Anannotated checklist of the aquatic macrophytes of theShire River, Malawi, with reference to potential aquaticweeds. Kirkia, 13, 125–142.
Chimatiro, S. and Mwale, D. 1998. A participatory ruralappraisal study on the socio-economic impact of waterhyacinth on local communities of the lower shire area,Malawi. Report to Department for InternationalDevelopment Renewable Natural Resources KnowledgeStrategy programme. Malawi Fisheries Department andCAB International, Kenya, 101p.
Crossley, R. 1980. High levels of Lake Malawi. Universityof Malawi Staff Seminar Paper 2.
Harley, K.L.S 1991. Survey report of a project on exoticfloating African water weeds. Unpublished report to theCommonwealth Science Council, 151–177.
Hill, G., Day, R., Phiri, G., Lwanda, C., Njaya, F.,Chimatiro, S and Hill, M.P. 1999. Water hyacinthbiological control in the Shire River, Malawi. In: Hill,M.P., Julien, M.H. and Center, T.D., ed. Proceedings ofthe First IOBC Global Working Group Meeting for theBiological and Integrated Control of Water Hyacinth.Harare, Zimbabwe, 16–19 November 1998. PlantProtection Research Institute, South Africa, 39–50.
Julien, M.H. and Griffiths, M.W. 1998. Biological control ofweeds: a world catalogue of agents and their target weeds,4th ed. Wallingford, UK, CABI Publishing.
Julien, M.H. and Orapa, W. 1999. Structure andmanagement of a successful biological control project forwater hyacinth. In: Hill, M.P., Julien, M.H. and Center,T.D., ed., Proceedings of the First IOBC Global WorkingGroup Meeting for the Biological and Integrated Controlof Water Hyacinth. Harare, Zimbabwe, 16–19 November1998. Plant Protection Research Institute, South Africa,123–134.
Rother, J.A. and Twongo, T. 1999. Water hyacinth in theShire River: techniques for monitoring biodiversityimpacts and a baseline survey. Unpublished report toDFID/Government of Malawi/CABI for Lower Shirewater hyacinth biological control project.
Terry, P.J. 1991. Water hyacinth in the Lower Shire,Malawi, and recommendations for its control. Report of aconsultancy for ODA, 18 September–5 November 1991.
57
IMPECCA1: an International, Collaborative Program to Investigate the Development
of a Mycoherbicide for Use against Water Hyacinth in Africa
R. Bateman*
Abstract
The IMPECCA Programme has been established to develop a mycoherbicide for the control of water hyacinth,using fungal isolates that have been found in Africa. Such a mycoherbicide could replace the use of broad-spectrum herbicides, which are used routinely at present, but have caused concerns about contamination to fishand degradation of water quality. In addition, a mycoherbicide might be more compatible with the use of insectbiological control agents. The project will build upon existing studies of formulating water hyacinth fungi intomycoherbicides which have been carried out in Egypt and Zimbabwe, and expertise gained by CABIBioscience during the development and commercialisation of mycoinsecticides.
One of the key outputs will be the strengthening of technical capacity and linkages within African nationalprograms to undertake biological control of weeds. Scientists will carry out extensive exploration forpathogens of water hyacinth already present in Africa. These will be identified, characterised and assessed forsuitability as the basis for a mycoherbicide; these studies will include molecular and chemotaxonomicidentification of both fungal isolates and water hyacinth biotypes. Characteristics of a fungus isolate suitablefor mycoherbicide development include: high pathogenicity, acceptable host specificity, low mammaliantoxicity, and capacity for mass production and formulation. Candidate products will be laboratory and fieldtested for efficacy and compatibility with other (especially biological) control options. A water hyacinthmanagement strategy will be proposed appropriate for local needs.
WATER hyacinth is perhaps the most perniciousaquatic weed in the world. Water hyacinth is generallythe dominant plant when it occurs outside of its native
range and is capable of suppressing or eliminatingother species. It forms dense mats of vegetation inlakes and dams, and irrigation and flood channels,where it impedes boat traffic, increases eutrophicationand harbours the mosquito vectors of malaria,encephalitis and filariasis (Forno and Wright 1981).The problems are most severe in developing countries,where human activities and livelihoods are closelylinked to the water systems. Conventional methods ofcontrol rely mainly on mechanical/manual removaland chemical herbicides, which have generally beenfound to be inadequate and expensive measures toapply on a large scale. Herbicides have the added dis-
1. The International Mycoherbicide Programme forEichhornia crassipes Control in Africa (the IMPECCAprogramme) is funded by Danida (Danish InternationalDevelopment Assistance) through the Environment, Peaceand Stability Facility. This article contains the views of theauthor, which do not necessarily correspond to the views ofDanida.
* CABI Bioscience, Silwood Park, Ascot, Berks, SL5 7TA,UK. Presented by Dr Bateman on behalf of program staff.
58
advantage that they might have adverse environmentaleffects, and must be applied carefully and selectively.They also can interfere with or nullify the action ofbiological control agents present (Charudattan 1995).In recent years, attention has been given to usingnatural enemies to control the weed. Indeed, the onlylogical, long-term and sustainable solution to man-aging the weed is to employ an integrated approach,with special emphasis on biological control agents(Charudattan 1995).
As described in other chapters of these proceedings,biological control using insects such as Neochetinaspp. has been highly successful in reducing water hya-cinth populations in several countries in northern andcentral America and parts of Africa. The need for arange of effective management tools against this weed,including mycoherbicides, is illustrated by the water-ways where water hyacinth remains a problem despiteattempts to introduce insects for biological control(Hill and Olckers 2001; Charudattan 2001). Further-more, several studies have demonstrated possibleinsect–pathogen interactions combining to cause adecline in water hyacinth populations (Galbraith 1987;Charudattan et al. 1978; Kasno et al. 1999). Mycoher-bicides might therefore be used to augment othernatural enemies.
Stages Required in the Development of a Mycoherbicide
Several research groups have identified promisingmicrobial agents that might be used as biopesticides.However, this research has rarely resulted in the devel-opment of biopesticide products, which currentlyaccount for only a fraction of one percent of the globalcrop protection market (which is approximately US$3× 1010). Although there is a long history of research onmicrobial control agents, it is not always appreciatedthat obtaining an active isolate is only the beginning ofa series of activities necessary for implementing theuse of a new mycoherbicide (O’Connell and Zoschke1996). There are important issues to considerincluding: mass production (see e.g. Jenkins et al.1998), delivery systems and ‘laboratory to field’studies (Bateman 1998), strategies for use, registrationand commercialisation.
Although focusing on ‘downstream’ processes ofmycoherbicide development, the IMPECCA Pro-gramme will also carry out further surveys for patho-gens for a limited period. Several isolates have alreadybeen collected that will act as standard isolates for
comparison with newly collected material. The criteriafor the selection of fungi were that:
• they must have been isolated in Africa;
• they must be widely distributed across the regionsof Africa where water hyacinth is found; and
• they must be shown to have good pathogenicity towater hyacinth in laboratory or field studies.
In contrast, Evans and Reeder (2001) argue for a‘longer term’ approach to the use of microbial agents,in which isolates are sought at the weed’s centre oforigin, for release as classical biological agents. How-ever, the introduction of exotic pathogens is in itsinfancy and there is still a reluctance to the release ofpathogens as biological control agents in many coun-tries. As part of the inception phase of the IMPECCAProgramme, the scientific collaborators have identi-fied a short-list of fungal species that show mostpromise in their potential as mycoherbicides for waterhyacinth control in Africa. In order of preference, theyare:
1. Alternaria eichhorniae;
2. Acremonium zonatum;
3. Cercospora piapori/C. rodmanii;
4. Rhizoctonia solani and Alternaria alternata;
5. Myrothecium roridum.
Program of Activities and Progress to Date
Table 1 gives a possible sequence of activities formycopesticide development, as identified byIMPECCA Programme scientists. The followingmajor activities are in progress or are planned:
• Collection and preservation of pathogen isolates.Over 70 isolates have been collated by CABIBioscience, and the most promising specimens willbe accessioned at CABI Bioscience, UK. Isolatecollection is being carried out in strict observance ofthe International Convention on Biodiversity, and adocument has been prepared explaining IMPECCAProgramme policy.
• Identification and characterisation of specimens:
– morphological, molecular and chemotaxonomicidentification of fungal isolates,
– molecular identification of water hyacinthbiotypes.
Preliminary analyses of water hyacinth isolatesfrom Africa, India and South America indicate asurprising degree of genetic homogeneity (AlexReid, pers. comm.). However, we foresee the need
59
Table 1. Major elements of mycoherbicide development
Issues Process (may be repeated with new isolate) Sequence
Survey for indigenous pathogens 1
Pathogen isolation and characterisation IsolationAgreement on standard isolatesPreservation of isolatesIdentification Assessment of pathogenicityCharacterisation
222345
Working methods Formation of country networksDevelopment of models as research toolsIntellectual property issuesPolicy on publication
1112
Biology Key protocols: screening techniques/objectivesWater hyacinth biologyWater hyacinth–pathogen interactions/nutrient levelsPersistence, horizontal transmission
119
10
Quarantine issues 2
Host range determination For environmental safetyEfficacy
66
Mass production Laboratory (Petri dishes / bottles)Pilot plantCommercial
27
13
Toxicology Mammalian safety tests (isolate assessment)Mammalian safety tests (registration)Ecotoxicology (includes host range data)Post control impact (succession of water hyacinth with other species) and changes in oxygen demand
812
6–12
10
Delivery systems StorageFormulationApplication
678
‘Lab to field’ studies ‘Pre field trials’ for efficacyAssessment of synergism (e.g. with Neochetina or chemicals/stressors)
99–11
Field testing Small-scale field trialsLarge-scale field trials (operational effectiveness)
1012
Strategy for use 10
Socioeconomic assessment 10
Product identification Ideally one product (maximum 2) – registration is expensive! 11
Registration dossier 12
60
to characterise approximately 100 samples beforeany firm conclusions can be drawn.
• Pathogen screening and selection for furthermycoherbicide development on the basis of acombination of key criteria including;– virulence,– host range testing,– preliminary mammalian toxicity studies, and– an assessment of production and storage
characteristics.A series of standard operating procedures for
isolate collection, and assessment of virulence andhost range testing, have been agreed or are inpreparation. Preliminary observations confirmthose of Shabana et al. (1995) that isolates ofAlternaria eichhorniae appear to be the mostpromising. These and other useful techniques arebeing compiled in the form of IMPECCA TechnicalGuides.
• Development of a suitable delivery system (Fig. 1)that includes:– preparation of prototype formulations,– storage stability tests, and – selection of appropriate application equipment.
The use of oil formulations has shown greatpromise for enhancing the efficacy of fungal agentsthat show potential as insecticides, fungicides andherbicides. Perhaps most importantly, the need forhigh humidity is overcome: Amsellem et al. (1990)showed that invert emulsions eliminated the need
for a minimum inoculum threshold with Alternariacassiae and A. crassa. Such formulations (where oilconstitutes the continuous phase) are rather ‘userunfriendly’ being unstable and very viscous, so theuse of less viscous vegetable oil suspensionemulsions has been investigated by Auld (1993) andShabana (1997). Bateman et al. (2000) discussimportant criteria in the selection of applicationequipment for biopesticides, and argue that theequipment normally used for conventionalpesticides is often most appropriate. Rotaryatomisers fitted to aircraft for low volume sprayingof chemical herbicides are a common method forwater hyacinth control (Julien et al. 1999) andespecially suitable for the application of oil-basedformulations.
• Field testing and ecological evaluation. Apreliminary model has been developed thatdescribes water hyacinth phenology; this will laterincorporate (and attempt to interpret) data on theimpact of natural enemies and their interactions(Neils Holst, pers. comm.)
• Programme management, liaison with collaboratingpartners, donors and other projects; agreements,patents, subcontracts, promotion and publicity.
The IMPECCA Programme collaborates with theIOBC by publishing The Water Hyacinth Newsletter(editors: Rebecca Murphy and Martin Hill). Newsand progress will also be available on our WorldWide Web page: <http//:www.impecca.net>.
Strategy for use
Stability
Host-pathogeninteractions
Operationalpracticalities
Mass production
Formulation
Application
Packaging andstorage
Figure 1. Delivery systems: successful biopesticide development is most likely tobe brought about by considering all the various technical aspects ofproduct development at an early stage. There are often importantlinkages between the mass production, storage and packaging processes.Formulation and application are likewise interdependent and governedby appropriate strategies for use of a microbial agent in the field.
61
Collaborators
The main investigators of the IMPECCA Programmeinclude:
1. CABI Kenya: Ignace Godonou, ChristiaanKooyman
2. CABI UK: Roy Bateman, Carol Elison, HarryEvans, Jane Gunn, Jeremy Harris, Nina Jenkins,Belinda Luke, Rob Reeder, Alex Reid, Sue Paddon,Emma Thompson
3. Danish Institute of Agricultural Science: NielsHolst (supervising PhD students: Sander Bruun andJohn Wilson)
4. Department of Research and Specialist Services,Zimbabwe: Bellah Mpofu, Lawrence Jasi
5. International Institute for Tropical Agriculture,Benin: Jürgen Langewald, Fen Beed
6. Plant Protection Research Institute, South Africa:Alana den Breeÿen, Cheryl Lennox
7. University of Mansoura, Egypt: Mahmoud Zahran,Mahomed El-Demerdash, Fathy Mansour, YasserShabana (from August 2001), Abdel-Hamid Khedr,Gamal Abdel-Fattah, Ibrahim Mashaly, S. El-Moursy
At the time of writing, the program has also estab-lished linkages with representatives from Kenya,Malawi, Rwanda, Tanzania and Uganda.
References
Amsellem, Z., Sharon, A., Gressel, J. and Quimby, P.C.1990. Complete abolition of high inoculum threshold oftwo mycoherbicides (Alternaria cassiae and A. crassa)when applied in invert emulsion. Phytopathology, 80,925–929.
Auld, B.A. 1993. Vegetable oil suspension emulsionsreduce dew dependence of a mycoherbicide. CropProtection, 12, 477–479.
Bateman, R.P. 1998. Delivery systems and protocols forbiopesticides. In: Hall, F.R. and Menn, J., ed.Biopesticides: use and delivery. Totowa, NJ, USA,Humana Press, 509–528.
Bateman, R.P., Matthews G.A. and Hall, F.R. 2000.Ground-based application equipment. In: Lacey, L. andKaya, H., ed. Field manual of techniques for theevaluation of entomopathogens. Netherlands, Kluwer,77–112.
Charudattan, R. 1995. Pathogens for biological control ofwater hyacinth. In: Charudattan, R., Labrada, R., Center,T.A. and Kelly-Begazo, C., ed., Strategies for waterhyacinth control. Report of a Panel of Experts Meeting 11–14 September, Fort Lauderale Florida, USA, 189–196.
Charudattan, R., Perkins, B.D. and Littell, R.C. 1978.Effects of fungi and bacteria on the decline of Arthropod-damaged water hyacinth (Eichhornia crassipes) inFlorida. Weed Science, 26, 101–107.
Charudattan, R. 2001. Biological control of water hyacinthby using pathogens: opportunities, challenges, and recentdevelopments. These proceedings.
Evans, H.C. and Reeder, R.H. 2001. Fungi associated withEichhornia crassipes (water hyacinth) in the upperAmazon basin and prospects for their use in biologicalcontrol. These proceedings.
Forno, I.W. and Wright, A.D. 1981. The biology ofAustralian weeds. 5. Eichhornia crassipes (Mart.) Solms.Journal of the Australian Institute of AgriculturalScience, 47, 21–28.
Galbraith, J.C. 1987. The pathogenicity of an Australianisolate of Acremonium zonatum to water hyacinth, and itsrelationship with the biological control agent, Neochetinaeichhorniae. Australian Journal of Agricultural Research,38, 219–229.
Hill, M.P. and Olckers, T. 2001. Biological controlinitiatives against water hyacinth in South Africa:Constraining factors, success and new courses of action.These proceedings.
Jenkins, N.E., Hefievo, G., Langewald, J., Cherry, A.J. andLomer, C.J. 1998. Development of mass productiontechnology for aerial conidia for use as mycopesticides.Biocontrol News and Information, 19, 21–31.
Julien, M.H., Griffiths, M.W. and Wright, A.D. 1999.Biological control of water hyacinth. Canberra,Australian Centre for International AgriculturalResearch, Monograph No. 60, 87 pp.
Kasno, Sunjaya, Putri A.S.R., Dhamaputra O.S. andHandayani H.S. 1999. Integrated use of Neochetinabruchi and Alternaria eichhorniae in controlling waterhyacinth. Biotropia, 13, 1–17.
O’Connell, P.J. and Zoschke, A. 1996. Limitations to thedevelopment and commercialisation of mycoherbicidesby industry. 2nd International Weed Control Congress,Copenhagen, 1189–1195.
Shabana, Y.M. 1997. Vegetable oil suspension emulsionsfor formulating the weed pathogen (Alternariaeichhorniae) to bypass dew. Zeitschrift für Pflanzenk-rankheiten und Pflanzenschutz, 104, 239–245.
Shabana, Y.M., Charudattan, R. and Elwakil, M.A. 1995.Identification, pathogenicity and safety of Alternariaeichhorniae from Egypt as a bioherbicide agent of waterhyacinth. Biological Control, 5, 123–135.
62
Fungi Associated with Eichhornia crassipes (Water Hyacinth) in the Upper Amazon Basin and
Prospects for Their Use in Biological Control
H.C. Evans and R.H. Reeder*
Abstract
Surveys were undertaken in 1998 and 1999 in the upper Amazon basin of Ecuador and Peru to collect andcatalogue the mycobiota associated with water hyacinth in the river and lake systems. The results indicate thatthree groups of fungi, which occupy distinct niches on the plant, can be delimited: biotrophic fungi, colonisinggreen leaf tissue, often without significant visible symptoms (e.g. Didymella and Mycosphaerella);necrotrophic fungi, causing prominent leaf lesions (e.g. Leptosphaeria, Colletotrichum, Myrothecium,Phaeoseptoria and Stagonospora); and fungi associated with and isolated from petioles previously invaded bycoevolved insect natural enemies, such as Taosa and Thrypticus spp. (e.g. Acremonium, Cephalosporiospsis,Cylindrocarpon, Cylindrocladium and Stauronema). Some of these represent new host records, as well asundescribed taxa. A re-analysis of the mycobiota associated with water hyacinth worldwide reveals that mostof the records originate from the USA and the Palaeotropics, where the plant is a major invasive species, andwhere, as a consequence, most research on its control has been concentrated. Fungal genera such as Alternariaand Cercospora, which traditionally have been favoured as biocontrol agents, seem to be absent or rare on E.crassipes in the Upper Amazon.
Introduction
EICHHORNIA crassipes (Mart.) Solms is native to theNeotropics but its precise centre of origin remainsspeculative. Based on style morphology, it has beensuggested that the area of greatest genetic diversity liesin Amazonia (Barrett and Forno 1982); with naturalspread from these to other regions of the South Amer-ican continent, and human-vectored introductions intothe Caribbean and Central and North America. Para-doxically, a search of the literature and unpublishedherbarium records reveals that few fungi have beenreported on water hyacinth in South America. Forexample, a detailed survey of the fungal pathogens
associated with this host in the Brazilian State of Riode Janeiro yielded only Cercospora piaropi, comparedwith four species recorded on the closely related Eich-hornia azurea (Swartz) Kunth. (Barreto and Evans1996). The same authors also compiled the worldwiderecords of the mycobiota collected on, or isolatedfrom, E. crassipes. A reanalysis of this amended list(Table 1) shows that of the 60 potential pathogensreported, 54 are from countries or regions where waterhyacinth is an undisputed alien invasive species, 36 ofwhich are exclusively Old World. Of the New Worldrecords, 18 are from the USA, 3 are from the Carib-bean or Central America, while only 2 have a SouthAmerican (ex Brazil) origin.
* CABI Bioscience, Silwood Park, Ascot, Berks. SL5 7TA,UK. Email: [email protected]
63
Table 1. Mycobiota recorded on Eichhornia crassipes, worldwide (amended from Barreto and Evans 1996)
Fungi Distribution
Ascomycotina and Deuteromycotina
Acremonium crotocigenum (Schol-Schwarz) W. Gams Australia (IMI 288071a)
Acremonium implicatum (Gilman & Abbott) W. Gams Australia (IMI 271067)
Acremonium sclerotigenum (F. & R. Moreau ex Valenta) W. Gams Sudan (IMI 284343)
Acremonium strictum W. Gams Australia (IMI 288318, 288319)
Acremonium zonatum (Sawada) W. Gams Australia, India, Pakistan, Panama, USA, Sudan
Alternaria alternata (Fr.) Keissler Egypt
Alternaria eichhorniae Nag Raj & Ponnappa Egypt, India, Thailand, USA, Kenya, Ghana, South Africa, Zimbabwe
Alternaria tenuissima (Nees ex Fr.) Wiltshire Hong Kong
Bipolaris urochloae (Putterill) Shoemaker Egypt (IMI 324728)
Bipolaris sp. USA, Brazil
Blakeslea trispora Thaxter Thailand
Cephalotrichum sp. USA
Cercospora piaropi Tharp India, Sri Lanka, USA
Cercospora rodmanii Conway USA–/India (IMI 329783), Nigeria (IMI 329211)
Chaetomella sp. Malaysia
Cladosporium oxysporum Berk. & Curt. Hong Kong–/Nigeria (IMI 333543)
Cochliobolus bicolor Paul & Parbery India (IMI 138935)
Cochliobolus lunatus (= Curvularia lunata) Nelson & Haasis Egypt (IMI 318639), India (IMI 162522, 242961), Sri Lanka (IMI 264391), Sudan (IMI 263783)
Coleophoma sp. Sudan (IMI 284336)
Curvularia affinis Boedijn USA
Curvularia clavata B.L. Jain India (IMI 148984)
Curvularia penniseti (M. Mitra) Boedijn USA
Cylindrocladium scoparium var. brasiliense Batista India
Didymella exigua (Niessl) Saccardo Trinidad, USA
Drechslera spicifera (Bainier) V. Arx Sudan
Exserohilum prolatum K.J. Leonard & E.G. Suggs USA
Fusarium acuminatus Ellis & Everhart Australia (IMI 266133)
Fusarium equiseti (Corda) Saccardo India–/Sudan (IMI 284344)
Fusarium graminearum Schwabe Australia (IMI 266133)
Fusarium moniliforme Sheldon Sudan (IMI 284342)
Fusarium oxysporum Schlechtendal Australia (IMI 288317)
Fusarium solani (Martin) Saccardo Australia (IMI 270062)
Fusarium sulphureum Schlechtendal India (IMI 297053)
Continued on next page
64
Most of the records in the exotic range, and espe-cially in the Palaeotropics, comprise a heterogeneousassemblage of generalist, opportunistic pathogens,with a minority group of apparently more specialisedspecies not yet recorded from the native range. Forexample, Cercospora piaropi was reported from Asia,Africa and North America only, before the aforemen-tioned survey in southern Brazil (Barreto and Evans
1996). Thus, Table 1 reflects the distribution of waterhyacinth research workers rather than the true coe-volved mycobiota. As Barreto and Evans (1996) con-cluded, the doubts and speculation surrounding thearea of origin or diversity of E. crassipes need to beresolved and addressed in order to open the way formore targeted and, potentially, more meaningfulsurveys for exploitable natural enemies.
Fusidium sp. South Africa (IMI 318345)
Gliocladium roseum Bainier Australia (IMI 278745)
Glomerella cingulata (Stonem) Spauld & Schrenk Sri Lanka (IMI 264392)
Helminthosporium sp. Malaysia
Leptosphaeria eichhorniae Gonzales Fragoso & Ciferri Dominican Rep., Panama
Leptosphaerulina sp. USA
Memnoniella subsimplex (Cooke) Deighton USA
Monosporium eichhorniae Sawada Taiwan
Mycosphaerella tassiana (De Notaris) Johanson USA
Myrothecium roridum Tode ex Fr. India, Philippines Thailand–/Burma (IMI 79771), Malaysia (IMI 277583)
Pestalotiopsis adusta (Ellis & Everhard) Steyaert Taiwan–/Hong Kong (IMI 119544)
Pestalotiopsis palmarum (Cooke) Steyaert India (IMI 148983)
Phoma sorghina (Saccardo) Boerema, et al. Sudan–/Australia (IMI 288313, 288311, 288312, 288315, 333325)
Phoma sp. USA
Phyllosticta sp. Nigeria (IMI 327627, 327628)
Spegazzinia tessarthra (Berk. & Curt.) Saccardo Sudan 284335
Stemphylium vesicarium (Wallroth) E. Simmons USA
Basidiomycotina
Doassansia eichhorniae Ciferri Dominican Rep.
Marasmiellus inoderma (Berk.) Singer India
Mycoleptodiscus terrestris (J.W. Gerdermann) Ostazeki USA
Rhizoctonia oryzae-sativae (Sawada) Mordue Australia (IMI 289087)
Rhizoctonia solani Kuhn India, Panama, Thailand and USA
Rhizoctonia sp. India, USA
Thanatephorus cucumeris (Frank) Donk China, Taiwan–/India (IMI 3075)
Tulasnella grisea (Raciborski) Saccardo & Sydow Indonesia (Java)
Uredo eichhorniae Gonzales Fragoso & Ciferri Argentina, Brazil, Dominican Rep.
Chromista
Pythium sp. USA
aInternational Mycological Institute isolate reference number
Table 1. (Cont’d) Mycobiota recorded on Eichhornia crassipes, worldwide (amended from Barreto and Evans 1996)
Fungi Distribution
65
The Surveys
Strategy employed
The Amazon basin, and specifically AmazonianBrazil, is most frequently cited as the probable centreof origin of E. crassipes (Harley 1990; Holm et al.1991). However, ad hoc pathology surveys along thelower Amazon and its tributaries in the early 1990s, inthe vicinities of Belém (Pará State) and Manaus (Ama-zonas State) yielded few fungi of interest (H.C. Evansand R.W. Barreto, pers. obs.). This led to speculationthat perhaps the true origin lay further south in thegreat basins of the Paraná or São Francisco rivers (Bar-reto and Evans 1996), particularly since the earliestrecord of the plant was from the Rio São Francisco(Seubert 1847). Nevertheless, an exploratory surveyalong this river in 1996 failed to find any new orexploitable pathogens (R.W. Barreto and H.C. Evans,unpublished data). The only major area in SouthAmerica for which there were no natural enemyrecords, and hence in which no surveys appeared tohave been conducted, is the northwestern region; spe-cifically, the upper Amazon basin, which comprises aconfluence of many river systems and interlinked orisolated lakes or ‘cochas’. It was hypothesised that insuch ecosystems, natural enemies of water hyacinthmay have coevolved in isolation and, as the plantspread naturally down the Amazon to reach theAtlantic and the other river systems of South America,these natural enemies were filtered out, especiallythose with poor survival or dispersal strategies. Thus,the biota associated with E. crassipes in the lowerAmazon basin and elsewhere may be depauperatecompared with that in the Upper Amazon, some 5500–7000 km upriver. The theory was put to the test, ini-tially by opportunistic surveys, followed-up later by amore organised collecting trip, in the upper Amazonbasin of both Peru (in Oct. 1988 and May 1999) andEcuador (in May and Sept. 1999, and May 2000).
Collecting and isolation
Collecting was done using motorised canoes, trav-elling down the Napo River in Ecuador from the portof Coca, and up the Amazon River from Iquitos inPeru and along the major feeder rivers of the Nanayand Marañon. In addition, a short survey was under-taken along the Ucayali River around the port ofPucallpa. Diseased leaves were collected and dried ina plant press for processing in the UK. In addition,plants were lifted and petioles, stems and roots exam-
ined for disease symptoms. Such fleshy material wasstored in waxed packets for later isolation in the UK.
Isolations were made either: directly from sporespresent on the diseased tissues, using a stereomicro-scope; or tissues were aseptically-dissected, surfacesterilised (30% hydrogen peroxide for 5 minutes) andrinsed several times in sterile distilled water. Allsamples were plated directly onto tap-water agar(TWA) or potato–carrot agar (PCA) containing anti-biotics (penicillin, streptomycin sulfate), and incu-bated at 25ºC, with a 12-hour black light regime tostimulate sporulation.
Results
Field assessment
The striking, and initially depressing, observation ofwater hyacinth populations in the rivers and lakes ofthe upper Amazon basin is that there is little visibleevidence to signify the presence of fungal pathogens,especially compared to E. crassipes in its exotic rangewhere patches of senescing or dying plants are notuncommon (caused by both abiotic and biotic factors).However, closer examination reveals that there is arange of fungal pathogens occurring on water hyacinth(see Table 2), and that these fungi fall into threegroups. Genera, such as Didymella and Myco-sphaerella, produce their discrete, black ascostromatasingly but abundantly in the still green leaf tissues and,thus, apart from some yellowing (chlorosis), symp-toms are cryptic. These species represent highly coe-volved or biotrophic fungi, living within the hostwithout seriously disrupting its physiology. Thesecond group includes fungi which belong to thegenera Colletotrichum, Leptosphaeria, Myrothecium,Phaeoseptoria and Stagonospora, and which causenecrotic leaf spots: some restricted and discrete (e.g.Colletotrichum); others spectacular, such as a promi-nent target spot (Leptosphaeria). However, it is onlywhen the plants are lifted, and the petioles examined,that the high incidence of disease becomes evident.Many petioles were attacked by species of Taosa (Dic-tyopharidae; Homoptera) and Thrypticus (Dolichopo-didae; Diptera), with their characteristic feeding andegg-laying patterns, and a significant proportion ofthese showed a positive association with fungalnecrosis, as evidenced by lesion development aroundand subsequent spread from the insect punctures. It isconsidered that these wounds permit the ingress ofboth specialist and opportunistic fungal pathogens intothe petiole, resulting in colonisation and invasion of
66
the stele, with decline or eventual death of the plantcaused, in part, by the actions of this third group offungi. There is a less clear association with the tunnelsof Neochetina larvae, although microorganismsreadily invade such damaged tissues. Interestingly,
there was no association of fungi with the feeding scarsof Neochetina adults on the leaves. The fungi isolatedfrom these tissues, excluding well-documented andubiquitous saprophytic species, are listed in Table 2.
Table 2. Mycobiota associated with Eichhornia crassipes in the Upper Amazon basin
Identification Country Associated tissue Isolate reference no.c
Ascomycotina and Deuteromycotina
Acremoniella sp. Peru Petiole –d
Acremonium sp. (New species) Peru Petiole 384422
Acremonium sp. (New species) Peru Petiole 384429
Acremonium sp.a Peru Petiole 384427
Asteroma sp. Peru Petiole 379974
Cephalosporiopsis sp. Peru Petiole –
Cephalosporium sp. Ecuador Leaf –
Chaetophoma sp. Ecuador Leaf –
Cochliobolus lunatus R.R. Nelson & F.A. Haasis Peru Petiole 379965
Cochliobolus pallescens (Tsuda & Ueyama) Sivan Peru Petiole 379971
Coniothyrium sp. Ecuador Petiole –
Curvularia sp. Ecuador Petiole
Fusarium sp. (New species) Peru Petiole 384418
Cylindrocladium sp.a Peru Petiole 384414
Fusarium poae (Peck) Wollenw. Peru Petiole 384424
Fusarium sacchari (E.J. Butler & Hafiz Kahn) W. Gams.
Peru Petiole 384423
Fusarium sp. (New species) Ecuador Petiole 384434
Fusarium sp. Peru Petiole –
Fusarium sp. Peru Petiole –
Fusarium sp.a Ecuador Petiole 384433
Gliocladium roseum Bainier Ecuador Petiole 384435
Gliocladium sp. Peru Petiole
Glomerella cingulata (Stoneman) Spauld. & H. Schrenk.
Brazilb Leaf 384437
Glomerella cingulata (Stoneman) Ecuador Leaf 384432
Glomerella cingulata (Stoneman) Peru Leaf 384416
Glomerella sp. Peru Leaf –
Hyphomycete sp. 1a Ecuador Petiole 384431
Hyphomycete sp. 2a Ecuador Petiole 384430
Hyphomycete sp. 3 Ecuador Petiole –
Hyphomycete sp. 4 Ecuador Petiole –
Hyphomycete sp. 5 with dictyochlamydospores (New species).
Peru Petiole 379967
Continued on next page
67
Laboratory assessment
An analysis of the fungi collected on, and isolatedfrom, diseased water hyacinth samples in the upperAmazon basin shows some notable differences fromthose fungi reported from other countries or regions(see Discussion). Notable among the Amazonianrecords are undescribed species of Acremonium (2spp.), Fusarium (2 spp.), Mycosphaerella (1 sp.) andprobably undescribed taxa, belonging to the genera
Phaeoseptoria, Stagonospora and Pseudocer-cosporella, since there are no previous records ofthese genera from E. crassipes. In addition, there arestill some tentative identifications which may repre-sent novel species and/or genera, and for which moretaxonomic inputs are awaited. In this context, of par-ticular interest is Hyphomycete sp. 5, which cannot beassigned to any known genus or indeed a taxonomicgroup. In culture, this fungus produces masses of
Idriella sp.a Peru Petiole 384417
Leptosphaeria sp. Brazilb Leaf –
Leptosphaeria sp.a Peru Leaf 384425
Leptosphaerulina sp. Peru Leaf 379972
Mycosphaerella sp. (New species) Peru Leaf 384426
Myrothecium verrucaria (Alb. & Schwein.) Ditmar Peru Leaf 379973
Myrothecium sp. Brazilb Leaf –
Phaeoseptoria sp. Peru Leaf 379966
Phoma chrysanthemicola Hollós Peru Petiole 384421
Phoma leveillei Boerema. & Bollen Ecuador Petiole –
Phoma section Peyronellaea (Goid. ex Togliani) Boerema
Peru Petiole 384420
Phoma sp.a Brazilb Petiole 384436
Phoma sp.a Peru Petiole 384428
Phoma spp. Ecuador Petiole –
Phoma spp. Peru Petiole –
Pseudocercosporella sp. Peru Leaf 384415
Sarocladium sp. Peru Petiole –
Stagonospora sp. Peru Leaf –
Stauronema sp.a Peru Petiole 384419
Basidiomycotina
Basidiomycete sp. 1 Peru Petiole –
Basidiomycete sp. 2 Peru Petiole –
Basidiomycete sp. 3 Peru Petiole –
Rhizoctonia sp. Ecuador Petiole –
Rhizoctonia sp. Peru Petiole –
Thanetophorus sp. Peru Petiole –
a Preliminary identification awaiting confirmation from CABI Bioscience, International Mycological Institute (Egham) or Centraalbureau voor Schimmelcultures (Baarn, Netherlands).b Recent survey along the Xingu River (Pará).c International Mycological Institute Herbariumd – = not yet accessed in collections.
Table 2. (Cont’d) Mycobiota associated with Eichhornia crassipes in the Upper Amazon basin
Identification Country Associated tissue Isolate reference no.c
68
hydrophobic, greenish-grey resting bodies (‘scle-rotia’) within which are produced more thick-walledresting structures or dictyochlamydospores. It can behypothesised that the ‘sclerotia’ are adapted forfloating and for dispersal of the resting spores,perhaps attaching to the leaves and petioles of waterhyacinth plants, but the rest of the fungal life-cycle,and specifically its invasion of the host, remainshighly speculative.
Greenhouse assessment
Several of the Acremonium species and other verti-cillioid Hyphomycetes have been screened on waterhyacinth plants (ex Africa) in a quarantine greenhousefacility in the UK. Only one species (Cephalospori-opsis) has demonstrated high pathogenicity; causing aspreading necrosis and death of inoculated,unwounded petioles. Clearly, more in-depthscreening, particularly with and without wounding (tosimulate insect attack), is necessary before the poten-tial of these Amazonian fungi as biocontrol agents ofE. crassipes can be properly evaluated.
Discussion
These essentially preliminary surveys demonstratethat there is a rich mycobiota associated with E.crassipes in the upper Amazon basin. Moreover, fewof these species share a common link with the myco-biota recorded in other regions or countries where theplant is an alien invasive species. For example, of theubiquitous pathogens which have been targeted andassessed as biocontrol agents of water hyacinth, onlyMyrothecium roridum has been found in both situa-tions. This suggests that other common taxa andpotential biocontrol agents such as Alternaria eich-horniae, Acremonium zonatum and Cercospora rod-manii (= C. piaropi), which have been recorded duringroutine surveys in the USA (Freeman et al. 1974),South Africa (Morris et al. 1999) and India (Evans1987), are altogether absent or rare on E. crassipes inthe upper Amazon.
Indeed, the origins and, in particular, the originalhost(s) of A. eichhorniae can only be speculated upon.Since its description on E. crassipes in India (Nag Rajand Ponappa 1970), it has been recorded from variouscountries in Africa, as well as from Egypt and theUSA (Table 1). However, pathogenicity tests in thelatter two countries showed contrasting results, withvirulent strains being reported in Egypt (Shabana etal. 1997) but only weakly pathogenic isolates in the
USA (Freeman et al. 1974). This fungus is alsoregarded as a weak pathogen in South Africa (Morriset al. 1999), although virulent strains have recentlybeen found in both East and West Africa (Bateman2001). Nag Raj and Ponappa (1970) reported that A.eichhorniae has a narrow host range, at least in thetests that were conducted, and attacked only a relatedmember of the Pontederiaceae (Monochoria vaginalisPers.). If E. crassipes is South American in origin, andif, as the present survey suggests, A. eichhorniae is notpresent in South America (or at least the upperAmazon), then what is its natural host range? A con-firmed record of this species on Bupleurum falcatumL. (Umbelliferae) from Germany (Evans 1987) onlyfuels the speculation.
Despite the spectacular success of Neochetinaweevils as classical biocontrol agents in a number ofcountries or regions, such control has not alwaysproven to be sustainable or universal, and hence thesearch for, and assessment of, other arthropod naturalenemies still continues apace (Cordo 1999; Hill andCilliers 1999). The essentially provisional resultsreported here indicate that new and potentiallyexploitable fungal pathogens can be found in theupper Amazon basin. The case for this being thecentre of origin or diversity of E. crassipes, therefore,has been strengthened but there are still some anom-alies. For instance, two biotrophic fungi, the rustUredo eichhorniae and the smut Doassansia eichhor-niae, which were described by the great Italianmycologist R. Ciferri on water hyacinth in theDominican Republic in the 1920s (Evans 1987), werenot found during the Amazonian surveys. A rust,however, was common on E. azurea in the same hab-itats. If these represent coevolved taxa, then thiswould suggest that the Caribbean is the true centre oforigin. Nevertheless, to support this conclusion, thehost range of the rust requires clarification, and theidentification of the smut needs to be verified. Unfor-tunately, a recent survey in the Dominican Republicfailed to locate either of these natural enemies (R.W.Barreto, pers. comm.).
It is relevant here to ask whether or not classicalfungal biocontrol agents could make a useful additionto the armoury to be deployed against E. crassipes inits exotic range, and, if so, is it an acceptable strategy?A judgment cannot yet been made on this questionsince classically introduced fungi have never beenused for management of water hyacinth in most of thecountries affected by the weed, where the introductionof exotic pathogens as biocontrol agents is still viewedwith considerable scepticism (Evans 2000). However,
69
based on recent results in South Africa and Australia(Evans 2000), this can be a potentially highly suc-cessful strategy and one which can be approachedfrom three possible directions.
Firstly, the traditional classical approach can beadopted, involving the release of a virulent, coevolvedfungal agent producing abundant inoculum withhighly efficient dispersal and survival mechanisms,such as a rust or smut. However, from the mycobiotadocumented so far (Tables 1 and 2), there is no indica-tion that a suitable candidate has been found. In fact,the majority of fungi recorded in the upper Amazonare either poor sporulators (e.g. Didymella andMycosphaerella), producing relatively few, delicateascospores; or possess slime-spores (conidia) whichare adapted for short-distance, rain-splash dispersalonly (e.g. Colletotrichum, Acremonium, Fusarium,Phaeoseptoria and Stagonospora). This restricted dis-persal ability may account for the fact that they appearnot to have spread with the plant during its migrationfrom the headwaters of the Amazon. The exploitationof such fungi as ‘classical’ mycoherbicides could beconsidered, in which the strategy would be to spot-spray rather than blanket-spray, allowing for naturalspread (rain or water-splash) within contiguous popu-lations, and perhaps a single application, rather thanrepeated doses, relying on the specialised survivalpropagules to ensure carryover and thus provide long-term or sustainable control.
However, perhaps the most potent use of these fungiwould be in conjunction with insects, as recommendedby Charudattan et al. (1978), and there is evidencefrom the current surveys that there is a close associa-tion between certain fungal species listed in Table 2and insect natural enemies such as Taosa and Thryp-ticus species. Indeed, an analysis of the early datarelating to prickly pear control in Australia, revealsthat success was achieved through a combination ofCactoblastis cactorum and the introduction of exoticmicroorganisms (Mann 1970).
Acknowledgments
These surveys were conducted within an AlternativeCrops Project funded by USDA-ARS through EricRosenquist (Office of International Research Pro-grams). The assistance of Djami Djeddour and SarahThomas during the surveys is gratefully acknowledged.
References
Barreto, R.W. and Evans, H.C. 1996. Fungal pathogens ofsome Brazilian aquatic weeds and their potential use inbiocontrol. In: Moran, V.C. and Hoffmann, H., ed.,Proceedings of the IX International Symposium onBiological Control of Weeds, Stellenbosch, South Africa,19–26 January 1996. University of Cape Town, SouthAfrica, 121–126.
Barrett, S.C.H. and Forno, I.W. 1982. Style morphdistribution in New World populations of Eichhorniacrassipes (Mart.) Solms-Laubach (water hyacinth).Aquatic Botany, 13, 299–306.
Bateman, R. 2001. IMPECCA: an international,collaborative program to investigate the development of amycoherbicide for use against water hyacinth in Africa.These Proceedings.
Charudattan, R., Perkins, B.D. and Littell, R.C. 1978.Effects of fungi and bacteria on the decline of arthropod-damaged water hyacinth (Eichhornia crassipes) inFlorida. Weed Science, 26, 101–107.
Cordo, H.A. 1999. New agents for biological control ofwater hyacinth. In: Hill, M.P., Julien, M.H. and Center,T.D., ed., Proceedings of the First IOBC Global WorkingGroup Meeting for the Biological and Integrated Controlof Water Hyacinth, Harare, Zimbabwe, 16–19 November1998. Pretoria, South Africa, Plant Protection ResearchInstitute, 68–74.
Evans, H.C. 1987. Fungal pathogens of some subtropicaland tropical weeds and the possibilities for biologicalcontrol. Biocontrol News and Information, 8, 7–30.
— 2000. Evaluating plant pathogens for biological controlof weeds: an alternative view of pest risk assessment.Australasian Plant Pathology, 29, 1–14.
Freeman, T.E., Zettler, F.W. and Charudattan, R. 1974.Phytopathogens as biocontrols for aquatic weeds. PANS,20, 181–184.
Harley, K.L.S. 1990. The role of biological control in themanagement of water hyacinth, Eichhornia crassipes.Biocontrol News and Information, 11, 11–22.
Hill, M.P. and Cilliers, C.J. 1999. A review of thearthropod natural enemies, and factors that influencetheir efficacy, in the biological control of waterhyacinth, Eichhornia crassipes (Mart.) Solms-Laubach(Pontederiaceae), in South Africa. African EntomologyMemoir, No. 1, 103–112.
Holm, L.G., Plucknett, D.L., Pancho, J.V. and Herberger,J.P. 1991. The world’s worst weeds. Florida, KriegerPublishing Company, 609p.
Mann, J. 1970. Cacti naturalised in Australia and their control.Brisbane, Queensland Department of Lands, 128p.
70
Morris, M.J., Wood, A.R. and den Breyen, A. 1999. Plantpathogens and biological control of weeds in SouthAfrica: a review of projects and progress during the lastdecade. African Entomology Memoir, No. 1, 129–137.
Nag Raj, R.R. and Ponappa, K.M 1970. Blight of water-hyacinth caused by Alternaria eichhorniae sp. nov.Transactions of the British Mycological Society, 55,123–130.
Seubert, M. 1847. Pontederiaceae. In: Martius, C.P., ed.Flora Brasiliensis, 3, 85–96.
Shabana, U.M., Baka, Z.A.M. and Abdel-Fattah, G.M.1997. Alternaria eichhorniae, a biological control agentfor water hyacinth: mycoherbicidal formulation andphysiological and ultrastructural host responses.European Journal of Plant Pathology, 103, 99–111.
71
A Water Hyacinth Resource Manual
G. Hill* and R. Day†
PROBLEMS associated with water hyacinth infestationsare well documented throughout those parts of the devel-oping world where the weed has become a menace.Annual recurrent costs associated with water hyacinthglobally have been estimated to more than US$100m(Joffe and Cooke 1997). Solutions to the problem ofwater hyacinth infestations are many and varied, anddepend upon the particular situation in which the weedhas appeared, the level of infestation, and the kinds ofcommunities and facilities that are being affected, but areusually divided into three categories: biological control,chemical control and physical control. In addition tothese, a wide range of processes and applications hasbeen developed for the utilisation of water hyacinth.
The extensive published and unpublished literature onwater hyacinth control and utilisation is characterised bya dearth of information in two key areas: 1. control deci-sion-making and the integration of control options, and 2.the integrating of control and utilisation. To overcomethese obstacles to the effective control and utilisation ofthe weed, we propose to develop a comprehensive,authoritative and practical water hyacinth resourcemanual. The manual will be targeted principally at deci-sion-makers and project implementers in developingcountries, but would have information which would be ofvalue and interest to anyone involved in water hyacinthcontrol and utilisation. The contents of the resourcemanual will include:• details of all currently available control options;• a guide to weed utilisation;• guidelines for integration of different control options;• guidelines for integration of control measures with
utilisation;• management decision aids for different infestation
scenarios;
• how to design an integrated control and utilisationproject; and
• a comprehensive directory of resources andinformation sources.The project is a collaborative undertaking involving
several organisations (CABI Bioscience, the Interna-tional Union for the Conservation of Nature (IUCN),Anamed, Clean Lakes Inc., the United Nations Environ-ment Programme) and forms part of the work program ofthe IUCN Africa Regional Office and the Global Inva-sive Species Group (GISP). It will link directly with andprovide information for the IUCN initiative ‘Wetlandsand Harmful Invasive Species in Africa—Awareness andInformation’. A team of four technical editors, assisted bya professional editor, will produce the manual. The teamwill include specialists in physical, chemical and biolog-ical control, wetland management and utilisation of waterhyacinth. The editors will use an iterative (Delphi)process of consultation with a large group of specialistsworking on water hyacinth from around the world, on thecontent of drafts of the manual. It plans to consult widelyamongst the members of the IOBC Global WorkingGroup on Water Hyacinth. This will ensure that the con-tents and recommendations are as authoritative and com-plete as possible.
The plan is to have a first draft of the manual preparedby mid 2001, with a final publication date 9–12 monthsafter that. Further information can be obtained fromGarry Hill at <[email protected]>.
ReferenceJoffe, S. and Cooke, S. 1997. Management of water hyacinth
and other invasive aquatic weeds. Issues for the World Bank.Washington, DC, World Bank internal report, 36p.
* CABI Bioscience UK Centre, Silwood Park, Ascot,Berkshire, SL5 7TA, UK.
† CAB International Africa Regional Centre, PO Box 633,Village Market, Nairobi, Kenya.
72
Water Hyacinth Information Partnership for Africa and the Middle East
L.A. Navarro*
Abstract
A ‘water hyacinth information partnership’ is proposed as an information–communication mechanism tofacilitate timely decisions in cases of water hyacinth infestations across Africa and the Middle East. The ideaarose from a consultation of stakeholders across the region, which was supported by the InternationalDevelopment Research Centre in 1996–1997. The proposal responds to the finding that countries across Africaand the Middle East usually start to control water hyacinth too late, after infestations have reached crises levels,despite the availability of expertise within the region. The partnership is to serve the countries as a decision-support information–communication mechanism, making the region able to detect and respond early andcost-effectively to infestations of water hyacinth in its water bodies. Its mission is to facilitate communicationand exchange of information on water hyacinth among affected people, decision-makers, experts and donors,thereby contributing to control of the weed. It will serve its constituency by: facilitating their access toscientific information on water hyacinth, both biophysical and socioeconomic; raising awareness amongdecision-makers and leaders about the characteristics of the weed and of the implications for infested waterbodies and the people who depend on them; helping to identify and mobilise expertise and resources availablefor the control of water hyacinth within the region and globally; calling early attention to impending waterhyacinth infestations in water bodies of the region; and championing early and effective control efforts of theweed. The funding for and specific plans to install the partnership are still under discussion.
THE Water Hyacinth Information Partnership (WHIP)has been conceptualised as an information–communi-cation mechanism to alert communities and especiallydecision-makers concerned with water bodies ofAfrica and the Middle East (AME), including Egypt,Lebanon, Syria, Jordan, Palestine and Israel, that arefacing impending infestations of water hyacinth. Itwould also foster and facilitate quick reaction to thethreat by providing countries with timely information.
The vision is that of a region that is able to halt, andit is hoped, revert the spread of water hyacinth acrossits water bodies, and thereby prevent water hyacinth
from reaching costly crisis levels in any water body inthe region.
WHIP’s mission is, through the use of modern andmore traditional information–communication technol-ogies, to target and tap key sources of information andexpertise on water hyacinth and to mobilise decision-makers and to stimulate efforts to control the weed. Inthe longer term, the expectation is that WHIP wouldfoster and support the integrated management of waterbodies and their basins to diminish soil erosion andother sources of water pollution that favour the growthof aquatic weeds.
WHIP’s Origins and Rationale
The idea and concepts of WHIP emerged from a 1996–97 consultation of selected researchers, decision-
* Senior Program Specialist, Agricultural and NaturalResources Economics, International DevelopmentResearch Centre, P.O. Box 62084, Nairobi, KenyaEmail: [email protected]
73
makers, donors and community leaders concernedwith water hyacinth across AME. This consultationbegan in late 1996 with a survey of key informantsimplemented by a team of 5 expert consultants across29 countries of the region. These countries includedthose with the most experience of water hyacinth, suchas Benin, Egypt, Kenya, Nigeria, South Africa, Sudan,Tanzania, Uganda and Zimbabwe. The consultationended with a consultative workshop of water hyacinthexperts and stakeholders (‘Improving reaction to waterhyacinth in affected countries across Africa and theMiddle East; consultative workshop on the capabilityof communities, authorities and organisations to reactand handle problems of water hyacinth in the region),held in Nairobi, Kenya, in September 1997 (Navarroand Phiri 2000).
The results of the survey and consultation indicatedthat water hyacinth was present in all 29 countries sur-veyed and had reached crisis levels in 21 of them.
Water hyacinth entered AME in the late 1800s inEgypt. Its spread indicates that it also later enteredthrough other countries. The spread of the weed hasaccelerated and become critical since the 1980s.
Water hyacinth infestations have been worst in theintricately connected water bodies of eastern andsouthern Africa. The most recent hot spot, in terms ofcrisis water hyacinth infestation, has been Lake Vic-toria in East Africa.
The consultation also revealed that mechanical andlabour-intensive manual methods of water hyacinthcontrol have been the most commonly used in AME,despite their acknowledged higher costs. Chemicalcontrol was used successfully in earlier efforts tocontrol water hyacinth e.g. in Egypt, South Africa andZimbabwe. More recently, however, different coun-tries have grown wary of chemical control because ofconcerns for potential environmental damage, andhave shifted most of their interest to biological control,e.g. Lake Victoria. Countries such as Egypt havebanned the use of chemicals to control water hyacinth.
Finally, the consultation made clear that, whatevertype of control was used, organised and effectivecontrol of water hyacinth began only after infestationshad reached crisis levels in all known cases. This hap-pened even in cases where control has been deemedsuccessful, such as Benin, South Africa, Sudan, Zim-babwe and, most recently, Lake Victoria. The consul-tation also noted that the region now has sufficientexperience and expertise to manage water hyacinthinfestations.
Concern about Delayed Reaction to Water Hyacinth Infestations
Delayed reaction to infestations of water hyacinth,given available capabilities in the region, was the mainconcern expressed by the stakeholders surveyed. Suchconcern arises because of the speed with which waterhyacinth infestations can spread and the negative eco-nomic, social and environmental consequences ofwide water hyacinth infestations.
The cumulative cost of water hyacinth infestationfor countries in AME is estimated to run to billions ofdollars. In the recent crisis in Lake Victoria, some esti-mates indicated that water hyacinth covered at least40,000 ha at its peak, affecting the livelihoods of manyfishing and other riparian communities in Kenya, Tan-zania and Uganda. For example, at the end of 1997media agencies reported a 70% decline in economicactivities at the Kenyan port of Kisumu as a result ofwater hyacinth choking the port and fish-landinggrounds. Port Bell in Kampala was also closed forperiods as a result of water hyacinth mats. The waterhyacinth infestation in Lake Victoria has recededrecently, due to the release of two Neochetina weevilspecies.
Stakeholders consulted are aware that a quickerresponse would help to minimise the social, economicand environmental damage and costs of water hya-cinth infestations, and that a longer term strategy isalso needed. The longer term effort should foster andsupport a focus on the integrated management of thebasins around affected water bodies to control nutri-ents polluting the water and stimulating water hya-cinth growth. The intention is that WHIP wouldeventually include such concerns as part of its brief.
Reasons for Delays in Response to Water Hyacinth Infestations
The stakeholders identified institutional/organisa-tional, technical and financial reasons for the delays inthe responses to water hyacinth infestations.
Institutional/organisational reasons for delayedresponse were cited as the most common and wide-spread. These included lack of focused policies andinstitutional attention. Few countries have policiessuch as that in force in South Africa, which identifyand treat water hyacinth as a menace requiring publicmobilisation to control it. Usually there are too many,weak, uncoordinated and bureaucratic ‘water hyacinthunits’, with no clear mandate or leadership. Certainly,
74
there is a lack of early warning and information–com-munication mechanisms to inform decision-makersand quickly link them to sources of expertise andsupport when needs arise.
Technical reasons identified for delayed responseincluded a lack of well defined integrated control strat-egies. Studies of control efforts, even the successfulones, reveal reliance on improvisation, with little anal-ysis and use of existing experience. There is also anabsence of information on the spread and economic,social and environmental costs of water hyacinth withwhich to inform and alert the public, decision-makersand donors.
Generally, however, the experts that were consultedagreed that the region already has enough experience,knowledge and expertise to control any water hyacinthinfestation quickly, if these resources were mobilisedon time. There is also some experience in the use ofwater hyacinth but the approaches involved are not yetconsidered to be good control options.
Financial reasons for delayed response were oftencited, but not well defined. Although lack of funds wasusually cited as a matter-of-fact constraint, delayshave occurred even in cases where funds existed orinterested donors have been ready to help. In mostcases, there were other major reasons for the delay.
The Proposal
While delays in reaction to infestations with waterhyacinth were the main concern, the consultation alsoidentified an absence or tardy flow of existing infor-mation relating to water hyacinth among key playersas a major contributor to the problem.
In discussions during the survey and the closing con-sultative workshop, stakeholders identified the devel-opment and establishment of an information–communication mechanism to foster and supporttimely decisions and efforts to control water hyacinthusing regional capabilities, as the best immediateoption to help improve the existing situation. Theimprovement of the information–communication flowamong water hyacinth stakeholders, with a focus on thedecision-makers, was identified as the point of leastresistance and best option to start building on regionalstrengths to solve the ‘problem of water hyacinth’.
The initial proposal called for developing the con-cepts and blueprint for a ‘water hyacinth informationclearinghouse’. Participants at the Nairobi workshopin 1997 requested the International Development
Research Centre (IDRC) to further this proposal inconsultation with other donors and partners.
Water Hyacinth Information Partnership
IDRC, through its People Land and Water program,continued consulting with other donors and partners.These consultations indicated that the concept of aclearinghouse was considered too restricted orappeared to focus only on the contributions of scien-tific experts on water hyacinth. Since the intention wasto serve a wider constituency, a more inclusiveconcept was needed. Thus, the concept of an informa-tion partnership and the name of Water HyacinthInformation Partnership (WHIP) were adopted.
Vision and mission
WHIP has been conceptualised and is expected to bestructured and installed as a decision–support infor-mation–communication mechanism to serve the AMEregion, with the vision of making the region able todetect and respond rapidly and cost-effectively toinfestations of water hyacinth in the region’s waterbodies. As part of this, WHIP’s mission is to facilitatecommunication and the exchange of information onwater hyacinth among affected people, decision-makers, experts and donors, thereby contributing tocontrol of the weed and minimising its effects on thewell being and development of affected communitiesin AME.
Objective functions
As part of its mission, it is expected that WHIP willserve its constituency and especially its main users by:
• facilitating their access to biophysical andsocioeconomic information on water hyacinth;
• raising awareness among decision-makers andleaders about the characteristics of the weed and oftheir implications for infested water bodies and forthe people who depend on them;
• helping to identify and mobilise expertise andresources available for the control of water hyacinthwithin the region, and globally when necessary;
• calling early attention to impending water hyacinthinfestations in important water bodies of the region;and
• championing early and effective control efforts ofthe weed when and where needed.
75
Structure and organisation
WHIP will be constituted by the water hyacinthstakeholders—the beneficiary groups, and an infor-mation exchange and networking service—a servicegroup and its resources.
Water hyacinth stakeholder groupsThese groups will include:
• direct beneficiaries, including leaders, communitybased organisations, women and other groups incommunities affected by water hyacinth;
• decision makers—including policy-makers, publicofficers, managers, specialised research units andothers responsible for monitoring or control ofwater hyacinth;
• expert individuals and organisations, includingdocumentation centres, expert and research centresin universities and other units; and
• supporters, including donors, NGOs, the privatesector, the media, etc.
Information exchange and networking serviceAn information exchange and networking service
(IENS) will include the following personnel and facil-ities: • a coordinator—team leader; • secretarial, documentation and information–
communication technical staff support (the serviceteam); and
• housing facilities, equipment and materials,including a computer server and connectivity to theInternet and with stakeholders and partners.It will deliver its services through two types of
activities:• Core activities – in a permanent alert mode, which
will include:– updating of data on critical information needed or
which can be provided by different stakeholdergroups;
– updating databases on relevant data and availableliterature titles and their access;
– an awareness service to key stakeholders andgeneral information to all stakeholders;
– question-and-answer referral services; and – an Internet web site and discussion group
facilitation.• Special activities – in a championing and facilitating
mode when needs or opportunities arise:– organisation of workshops, seminars and short
courses;
– preparation or special packaging of trainingmaterials and tool kits—production of interactiveCD ROM, special web sites, etc.;
– development of specially targeted research andintervention proposals, and contributions to fund–raising; and
– management and implementation of specialstudies and projects.
ManagementIt is expected that the management of WHIP will be
in the hands of a steering committee that represents theassembly of stakeholders and is facilitated in its func-tions by the coordinator of IENS. The coordinatorIENS will be in charge of the day-to-day operationsand delivery of WHIP plans and services.
The WHIP steering committee will represent the‘assembly’ of stakeholders. It will be led by a chair-person and include a technical sub-committee and anexecutive sub-committee, to facilitate committeefunctions and support day-to-day operations.
The coordinator–team leader of IENS will havethe following functions and responsibilities:• executive secretary of the WHIP steering committee• lead the IENS unit and implement the WHIP work
program in consultation with stakeholders throughthe steering committee, including:– implementation and administration of the WHIP
programs and core activities;– preparation of annual work plans and budgets for
review and approval by the steering committee;– maintain contact with the steering committee
during plan implementation through the technicaland executive committees;
– maintain contact with and inform stakeholders ona continuous basis;
– champion and facilitate special activities,according to plans;
– facilitate steering committee meetings; and– facilitate fund-raising.
Estimated budget and issues to be resolvedAs result of the consultations and discussions to
date, the suggestion is to obtain support to install andoperate the WHIP for an initial period of five years.Given the level of activities and the cost of personnel,equipment and other support anticipated for the initialfive years, the estimated budget is US$1.5m.
The following issues remain to resolved :The host institution. Several institutions have
evinced interest in housing WHIP. The initial idea wasthat IDRC would house WHIP temporarily, allowing
76
time for discussions among the different stakeholdersto agree on a final location. Later ideas have suggestedthat the decision about where to house WHIP must betaken immediately. Thus accelerated consultations arerequired to reach agreement on this.
Water hyacinth only or invasive water weeds in gen-eral? A second interest emerging among stakeholdershas been to extend the coverage of WHIP to otherinvasive water weeds. This would seem to be a rationalextension of the coverage, but more discussion isneeded to make sure that such a move would not
obstruct the implementation of WHIP effort. The mainquestions relate to the implications of this idea onbudgetary and organisational matters, and on strate-gies for fund raising and allocation.
Reference
Navarro, L. and Phiri, G., ed. 2000. Water hyacinth in Africaand the Middle East. A survey of problems and solutions.International Development Research Centre, Ottawa,Canada. 120p. (http://www.idrc.ca/plan)
77
Can Competition Experiments Be Used to Evaluate the Potential Efficacy of New Water Hyacinth
Biological Control Agents?
T.D. Center*, T.K. Van*, and M.P. Hill†
Abstract
Two factors are of concern when considering a new biological control agent for introduction. The first is thesafety of the organism (i.e. its host specificity) and the second is the potential for the organism to control thetarget weed (i.e. its efficacy). Methods for evaluating safety before introduction are well known but scantattention has been paid to pre-release evaluation of the efficacy of the candidate organisms. This isunderstandable inasmuch as the agent’s performance depends on the presence or absence of density-dependentpopulation regulating factors that will differ between the donor area and the recipient country. However, thisis of less concern when the agent has already been introduced elsewhere, where it can be studied without theinfluence of density-dependent regulators. Experiments comparing the effectiveness of the new agent with thatof another, more widely known agent, can then be used to determine the relative value of the former with theknown impacts of the latter. Additive series analysis (inverse linear models) of competition between waterhyacinth and water lettuce as mediated by herbivory has been suggested as a means of judging the relativevalue of new agents. This approach is fraught with difficulties inasmuch as there will always be unknownfactors that affect the abundance of new agents (i.e. biotic resistance), but it could enable assessment of thepotential value of the proposed introduction and, in so doing, perhaps pre-empt the introduction of risky agentsthat provide little control value.
CLASSICAL biological control of a pestiferous non-native plant involves the deliberate introduction ofplant-feeding insects, mites, or phytopathogens (col-lectively called biological control agents, or herein,bioagents) from foreign sources to provide previouslymissing density-dependent regulation of the pestspecies in its adventive range. Typically, the bioagentis derived from within the native range of the pest andintroduced into a new area where control is needed.The safety of the introduced organism is of utmost
concern inasmuch as economically or ecologicallyimportant non-target plant species in the recipientregion may be at risk, and this risk escalates as moreand more agents are introduced. Thus, it is essential tointroduce the least number of species needed toprovide the control needed. In order to minimise thenumber of introductions, it would be useful to deter-mine beforehand which species, from among the cadreof potential bioagents available, would be the mosteffective. While this is often called for (Harris 1973),it is seldom done.
Techniques for determining the safety of a bio-agent, in terms of its fidelity towards the use of thetarget plant, consist mainly of bioassays of host specif-icity. These ‘host-specificity tests’ have a long historyof use and are very predictive (Pemberton 2000).
* USDA-ARS, Invasive Plant Research Laboratory, 3205College Ave., Fort Lauderdale, Florida 33314.
† Weeds Division, ARC, Plant Protection ResearchInstitute, Private Bag X134, Pretoria, 0001, South Africa.
78
However, standardised methods for evaluating thepotential impact of candidate bioagents are lacking.Harris (1973) and later Goeden (1983) attempted todevelop a scoring system based on specific attributesof the candidate agents. Both systems emphasised theamount of damage done to the plant on a per insectbasis or the capacity for population growth of theagent. Unfortunately, these tended to be ‘one size fitsall’ and failed to take into account the uniqueness ofeach weed–insect association. As a result, while theydo provide ‘rule-of-thumb’ guidelines, these scoringsystems are otherwise of limited usefulness. Amongother things, they fail to consider compensatory abili-ties and complementary characteristics of the varioustarget plants, which render them vulnerable (or not) todamaging effects of the agents. In other words, inorder to be effective, the damage done by the agentmust be directed towards the invasive attributes of theplant that enable it to dominate, so a useful scoringsystem must be tailored to each weed target. This isdifficult, at best, especially when, at the outset of aproject, so little is usually known about these agentsand the target plant.
There are few alternatives for directly assessing thevalue of a new, previously unused bioagent. Any suchappraisal must mimic a true biological control sce-nario in which the population increase of the agent isnot limited by density-dependent regulators, thus ena-bling their populations to attain greater densities thanthose normally found in the native environment. Suchassessments, which are best done under natural cir-cumstances, may be difficult to accomplish in thenative range of the bioagent, because of the presenceof density-dependent regulators that pre-empt buildupof the bioagent’s population and therefore fail to sim-ulate a true introduction scenario involving hyper-abundant bioagent populations. Furthermore, any suchassessment must be sensitive to subtle effects of thebioagents, so as not to disqualify those that mightprovide important, long-term effectiveness.
We propose direct experimentation to provide dataon the relative value of one agent compared to another.This does not resolve the difficulties involved in doingthe studies in the native area, but this approach is quitepossible when the agent has been previously intro-duced elsewhere and is being considered for introduc-tion into a new area. The mirid bug Eccritotarsuscatarinensis provides a useful example. It was firstintroduced into South Africa and is being consideredfor introduction into North America. Laboratory-
based host-specificity testing showed that it fed anddeveloped on pickerelweed, a valued North Americannative plant. Follow-up field studies in South Africarevealed that, while it might spill-over to pickerelweedwhen adjacent to heavily infested water hyacinth mats,it did little damage and did not colonise isolated pick-erelweed stands (Hill et al. 2000). Thus, it seems asthough this agent might, in fact, be safe to release inNorth America. However, the host-specificity dataclearly indicate that there is some risk to pickerelweed.Considering that pickerelweed is severely damaged bydrifting water hyacinth mats as well as by herbicidalcontrol operations directed against water hyacinth, thisrisk might be worth taking. The decision to release themirid must therefore weigh the potential damage topickerelweed against the benefit that it might provide.However, the effectiveness of the mirid is not yetknown. We are proposing to compare the effects of themirid with the effects of the better-known bioagent,the weevil Neochetina eichhorniae, on the competitiverelationship between water hyacinth and water lettuce.In so doing, we hope to determine whether the miridwould be more or less effective than the weevil and toquantify the difference.
The effects of the mirid are likely very subtle. It isa quite small insect that causes little damage per indi-vidual, which is neither overt nor easily quantified. Itfeeds on leaf surfaces by sucking plant juices, creatingbrownish patches that vary in extent and intensity(similar to spider mite damage). While this damagemay be debilitating to some degree, it does not seemlethal. In situations such as this, competition studiesmay be able to detect these subtle effects by measuringthe reduction of the plant’s competitive ability againstanother aggressive species. Pantone et al. (1989) pro-posed the use of additive series experiments analysedusing inverse linear models to evaluate the efficacy ofbioagents before release (although they did notaddress the aforementioned difficulties in doing thesestudies in the agent’s native range). They further dem-onstrated the utility of the method by detecting theeffects of a nematode on competition between the fid-dleneck weed and wheat. It thus occurred to us that thisapproach might be useful for determining the value ofthe water hyacinth mirid. We have used this modelpreviously (Van et al. 1998, 1999) to compare theinfluence of two hydrilla biological control agents andto investigate the effect of soil fertility on competitionbetween the two aquatic plants Hydrilla verticillataand Vallisneria americana.
79
Additive Series Competition Experiments and the Inverse
Linear Model
Pantone et al. (1989) provided a thorough explanationof additive series competition experiments and theapplication of the inverse linear model. Their papershould be consulted for details. Competition experi-ments involve planting mixtures of two plant speciesand, after a period of growth, measuring yield compo-nents of each species and comparing them betweenspecies. Additive series competition experimentsdiffer from replacement series competition experi-ments in that the total number of plants used for thetwo species varies as the mixture ratio increases (i.e. 3of species A vs. 0 of B for 3 total; 3 of A vs. 3 of B for6 total; 3 of A vs. 6 of B, or 9 total; etc.). In contrast,replacement series experiments use a constant totalnumber of plants while the ratio of the two varies.Pantone et al. (1989) used the mixtures given in Table1 in their additive series experiments
A control series was planted without nematodes anda duplicate second series (treatment) was planted andthe plots were inoculated with 106 fiddleneck gallnematodes (Anguina amsinckiae). Plants were har-vested after 5–6 months and average yield per plant (Y)was measured in terms of shoot dry weight, seednumber, and total seed biomass per plant.
Data were analysed using multiple linear regres-sions of the inverse of the yield component as thedependent variable and the planting density of wheatand fiddleneck as two independent variables as such:
1/Yf = afo + affdf + afwdw
1/Yw = awo +awwdw + awfdf
Here Yf is the average yield per plant for fiddleneck,Yw is the average yield per plant for wheat, df is theplanting density of fiddleneck, and dw is the planting
density for wheat. The coefficients aff and awwmeasure intraspecific competition of fiddleneck andwheat, respectively. The coefficient afw measures theinterspecific effect of wheat on fiddleneck yield, andthe coefficient awf measures the interspecific com-petitive effect of fiddleneck on wheat yield. The ratioaff/afw measures the effects of intraspecific competi-tion of fiddleneck on itself relative to the interspecificcompetition of wheat on fiddleneck. In other words, itequates the competitive effect of a single fiddleneckplant with the number of wheat plants that would beexpected to have an equivalent effect on fiddleneckyield (i.e. it takes x number of wheat plants to producethe same effect as a single fiddleneck plant on fiddle-neck yield). Likewise, the ratio aww/awf measures theeffect of wheat on wheat yield relative to the effect offiddleneck on wheat yield. The data can be graphicallyanalysed as a 3-dimensional surface response planefor each plant species in which the slope in one direc-tion represents the effect of the species own densityupon its yield (intraspecific competition) and the slopein the other direction represents the effect of the com-peting species (interspecific competition). It must beborne in mind that, because the inverse of thedependent variable is used, a higher value represents alower yield. Likewise, a steep slope represents astrongly reduced yield in response to increasing plantdensity. Results of one of the experiments conductedby Pantone et al. (1989) are presented in Table 2. Notethat increasing fiddleneck density strongly reducedfiddleneck yield per plant, as evidenced by the steepslope reflected in the coefficient aff, when nematodeswere absent. However, the effect of wheat on fiddle-neck yield per plant was slight. The ratio of the twovalues (aff/afw) indicates that the effect on fiddleneckyield of increasing the density of fiddleneck by asingle plant was equivalent to increasing the density ofwheat by 33 plants. When nematodes were present,however, the effects of the two species were similar, asreflected by the ratio of the two coefficients being nearunity.
The complementary analysis similarly indicates thatthe interspecific effect of fiddleneck on wheat yieldwas much greater than the intraspecific effect of wheaton itself. Increasing the density of wheat by one planthad the equivalent effect on wheat yield of adding 0.3fiddleneck plants. When nematodes were present thisincreased to 0.72 fiddleneck plants.
Recently, similar experiments have been done inFlorida (Van, unpublished data) to examine theeffects of the weevil N. eichhorniae on competitionbetween water hyacinth and water lettuce (Pistia
Table 1. The additive series planting ratios of wheatand fiddleneck used by Pantone et al. (1989)
Wheat Fiddleneck
:0 :20 :80 :160
0: 0:20 0:80 0:160
20: 20:0 20:20 20:80 20:160
80: 80:0 80:20 80:80 80:160
160: 160:0 160:20 160:80 160:160
80
stratiotes) (Table 3). In this example, without wee-vils, increasing the density of water hyacinth by oneplant produced 18.5 times the effect on water hya-cinth yield of increasing the water lettuce density byone plant. In other words, it required nearly 20 waterlettuce plants to produce the equivalent effect of asingle water hyacinth plant on water hyacinth yield.With weevils in the system, however, water hyacinthremained the superior competitor but its advantagewas reduced to less than 2 to 1. Likewise, withoutweevils an increase in water lettuce density of oneplant affected water lettuce yield by an amountequivalent to only 0.15 water hyacinth plants (or 7water lettuce plants were required to produce theeffect of 1 water hyacinth plant) but with weevilspresent this ratio increased to nearly unity.
Clearly, these analyses provide a useful way ofassessing the impact of a bioagent on two-speciescompetition, but can they be used to compare bioa-gents? The studies by Van et al. (1998) indicate thatthey can. They compared two hydrilla control agents interms of their effects on competition between H. vert-icillata and V. americana. They showed that, in theabsence of bioagents, intraspecific competition byHydrilla on itself was 8.3 times stronger than interspe-cific competition from Vallisneria. In the presence of
the leaf-mining fly, Hydrellia pakistanae, however,intraspecific and interspecific effects were nearlyequal (ahh/ahv = 1.3). The weevil Bagous hydrillaeproduced a much smaller shift in the competitivebalance (ahh/ahv = 7.6), which was not much differentfrom the control. As a result, one might conclude thatthe fly is nearly six times better than the weevil, interms of its ability to alter the competitive balancebetween these two plant species.
Given the positive results of these studies, we arenow comparing the two species of Neochetina(N. eichhorniae vs. N. bruchi) in terms of their abilityto alter the competitive relationship between waterlettuce and water hyacinth. The results are not yet in.This experiment involves 96 experimental units (8planting densities × 4 insect levels × 3 replicates). The8 planting densities (the minimum necessary) encom-pass factorial combinations of 0, 3, or 9 water hyacinthand water lettuce plants (minus the 0:0 combination).The insect treatments consist of N. eichhorniae alone,N. bruchi alone, both species together, or neitherspecies (as a control).
The logistics of setting up such a large experimenthave been difficult. Nevertheless, if this experimentproduces useful results, we are planning a similarexperiment to be conducted in South Africa to
Table 2. Regression coefficients from analyses of the effects of nematodes and plant density on reciprocals of thebiomass yields of wheat or fiddleneck (from Pantone et al. 1989)
Plant Treatment Regression coefficients
aff afw aff/afw aww awf aww/awf
Fiddleneck Control 8.24 0.25 33.0
Nematode 8.76 8.40 1.04
Wheat Control 4.97 16.4 0.30
Nematode 5.81 8.09 0.72
Table 3. Regression coefficients from multiple regression analyses of the impacts of weevils and plant density onthe reciprocal biomass yield of water hyacinth and water lettuce (from Van, unpublished data)
Plant Treatment Regression coefficients (× 10–3)
aww awl aww/awl all alw all/alw
Water hyacinth Control 0.943 0.051 18.5
Weevil 3.72 2.28 1.63
Water lettuce Control 9.41 62.1 0.15
Weevil 3.24 3.52 0.92
81
compare the mirid with N. eichhorniae. In so doing,we hope to quantify the effect of the mirid relative tothe effect of the weevils using the weevil as a standard.However, this involves another difficulty: how todetermine the numbers of each insect species to beused when two very different plant-feeding insects areinvolved. In the case of the two Neochetina species,this is not a problem. Both are about the same size andproduce the same type of damage. However, com-paring the chewing damage of the larger weevils withthe sap-sucking damage of the tiny mirid is anothermatter. Is it appropriate to merely use the same numberof each species, despite the size difference and the dis-parity in the type and amount of damage? Would it bebetter to introduce equivalent weights of both species?Obviously, it would be best to use a range of infesta-tion levels of each insect to measure the densities ofeach needed to produce equivalent effects, but the sizeof the experiment then becomes prohibitive. These andmany other questions must be resolved before pro-ceeding with plans for this experiment.
It is important to keep the limitations of theseexperiments in mind. First, cages are used and severaltypes of cage effects could lead to erroneous conclu-sions. Secondly, the experiments described aboveinclude only a few of the multitude of environmentalparameters that might affect the outcome of competi-tion. The effects of the insects might be compromised,for example, by high or low nutrients, but incorpora-tion of a nutrient treatment in the experiment designwould at least double the size to 192 experimentalunits in the case of the two-weevil experimentdescribed above. Thus, while it is important, if pos-sible, to retain the full additive series so as to producecomparable regression coefficients, it might not be
possible to answer all pertinent questions in thismanner. We are therefore considering additionalexperiments with varying nutrient levels but fixedcombinations of the two plant species for compari-sons between N. eichhorniae and N. bruchi. This isless desirable, but much more practical.
References
Goeden, R.D. 1983. Critique and revision of Harris’ scoringsystem for selection of insect agents in biological controlof weeds. Protection Ecology, 5, 287–301.
Harris, P. 1973. The selection of effective agents for thebiological control of weeds. The Canadian Entomologist,105, 1495–1503.
Hill, M.P., Center, T.D., Stanley, J., Cordo, H.A., Coetzee, J.and Byrne, M. 2000. Host selection by the water hyacinthmirid, Eccritotarsus catarinensis, on water hyacinth andpickerelweed: a comparison of laboratory and fieldresults. In: N.R. Spencer, ed., Proceedings of the 10th
International Symposium on Biological Control ofWeeds, 4–9 July 1999, Bozeman, Montana, 357–366.
Pantone, D.J., Williams W.A. and Maggenti, A.R. 1989. Analternative approach for evaluating the efficacy ofpotential biocontrol agents. 1. Inverse linear model. WeedScience, 37, 771–777.
Pemberton, R.W. 2000. Predictable risk to native plants inweed biological control. Oecologia, 125, 489–494.
Van, T.K., Wheeler, G.S. and Center, T.D. 1998.Competitive interactions between hydrilla (Hydrillaverticillata) and vallisneria (Vallisneria americana) asinfluenced by insect herbivory. Biological Control, 11,185–192.
—1999. Competition between Hydrilla verticillata andVallisneria americana as influenced by soil fertility.Aquatic Botany, 62, 225–233.
82
How Safe Is the Grasshopper Cornops aquaticum for Release on Water Hyacinth in South Africa?
I.G. Oberholzer and M.P. Hill*
Abstract
The grasshopper Cornops aquaticum is currently being considered as a natural enemy for water hyacinth inSouth Africa. Both the adults and the nymphs are very damaging to water hyacinth plants. The laboratory hostrange was determined through nymphal and adult no-choice trials. The test plants were selected onrelatedness to water hyacinth, similarity in habitat and on economic importance. Full nymphal developmentwas recorded on Heteranthera callifolia, Pontederia cordata (pickerelweed) and Canna indica (canna) underquarantine laboratory conditions. Pickerelweed and canna are introduced species and are potentially invasivein South Africa, and are therefore of no conservation concern. Of the other native African Pontederiaceae,Eichhornia natans supported development of the grasshopper nymphs, but the lack of emergent leaf materialsuggests that the plant will not sustain a population, and Monochoria africana did not support fulldevelopment of the nymphs. The adult females were not able to oviposit on the thin petioles of Heterantheracallifolia and only one eggpacket was recorded on Monochoria africana, suggesting these two species are notat risk. Results from the region of origin show that C. aquaticum is an oligophagous insect on thePontederiaceae family of plants, with a strong preference for water hyacinth. In South Africa we intend toconduct further nymphal and adult choice trials which will better represent the field situation to furtherquantify the risk to native Pontederiaceae.
WATER hyacinth is considered to be the most impor-tant aquatic weed in the world (Center 1994; Wrightand Purcell 1995). In South Africa, it was firstrecorded in the early 1900s. Since then the weed hasbecome invasive throughout southern Africa, mainlyas a result of human activities (Jacot Guillarmod1979). Attempts to control the weed have led to dif-ferent control options being developed, including her-bicidal control, mechanical control and biologicalcontrol. In South Africa, the biological controlprogram has been in place since 1974, with an inter-ruption of 8 years between 1977 and 1985 (Hill andCilliers 1999). In the course of the program, five
arthropod natural enemies were released against theweed: Neochetina eichhorniae, Neochetina bruchi,Orthogalumna terebrantis, Eccritotarsus catarinensisand Niphograpta albiguttalis. Even with these speciesreleased there is a perception that, in South Africa, thecorrect ‘suite’ of insects to biologically control theweed has not been introduced. As a result, additionalnatural enemies are being sought for control of waterhyacinth. In this paper we discuss the suitability ofCornops aquaticum, a grasshopper species, for releasein southern Africa.
Information from the Literature
Cornops aquaticum was identified by Perkins (1974) asbeing one of the most damaging insects associated withwater hyacinth in the plant’s region of origin. However,it appears that fears regarding this insect’s host specifi-
* Weeds Research Division, Plant Protection ResearchInstitute, Agricultural Research Council, Private BagX134, Pretoria, 0001, South Africa. Email: [email protected]
83
city have prevented it from being given serious consid-eration as a biological control agent for the weed.Silveira Guido and Perkins (1975) investigated thebiology and host specificity of C. aquaticum and foundthat, under laboratory starvation trials, it was able tofeed and develop on three species within the Commel-inaceae and also on the following species in the Ponte-deriaceae (Eichhornia azurea, Eichhornia crassipesand Pontederia cordata). Limited feeding, but nodevelopment, was recorded on rice and sugarcane.
Laboratory Determination of Biology
The grasshopper was collected from Brazil (1995),Trinidad and Venezuela (1996) and Mexico (1997)and imported into quarantine in South Africa. Theadult female inserts its eggs into the base of the petiole.According to Silveira Guido and Perkins (1975) theendophytic position of the egg packets provides mois-ture for development and the arenchyma tissue of thewater hyacinth petiole prevents excess water uptake.This might well be significant in the host specificity ofthe insect, as it appears as if the adult female has veryspecific ovipositioning requirements. These require-ments are unlikely to be present in plant speciesoutside of the Pontederiaceae. The egg cases are pro-duced inside a case of foamy substance, that functionsas a ‘plug’ to encapsulate the eggs. The ovipositionsite is identifiable by this plug, which the female usesto cover the oviposition hole. Eggs that were not ovi-posited within the plant tissue did not develop.
An incubation period of 25–30 days was recorded.Newly emerged nymphs begin to feed immediately onthe water hyacinth leaves. There are 6–7 instars (usu-ally 6) which range in length from 6–8 mm in the 1stinstar to 25–30 mm in the 6th instar. The adults arelong lived (55–110 days) and the females produce ahigh number of offspring: between 60 and 560. Theinsects are highly mobile and very damaging to waterhyacinth, both as adults and throughout the immaturestages.
Laboratory Host Specificity
Nymphal no-choice trials
Host range was determined through nymphal no-choice trials on 64 plants in 32 families, selected onrelatedness to water hyacinth, similarity in habitat andeconomic importance (Table 1). Details of the devel-
opment of C. aquaticum adults from no-choicenymphal starvation trials are presented in Table 2. Fivenewly hatched, first-instar nymphs were placed on eachof the test plant species. Feeding damage, nymphaldevelopment and mortality were recorded daily.
On the majority of species tested, no feeding wasrecorded and the nymphs died within the first week.Nymphal feeding was recorded on several speciesoutside the Pontederiaceae family. Some nibbling wasrecorded on rice and cabbage, but no development wasrecorded. A few nymphs developed to 2nd instar stageon radish, 3rd instar stage was reached on Nerine sp.(Amaryllidaceae) and 4th instar stage on Commelinaafricana and Murdannia simplex (both Commelin-iaceae). Complete nymphal development occurred onCanna indica, but the surviving number was low com-pared with survival on water hyacinth. Feeding anddevelopment were also recorded on pickerel weed, butnymphal survival was low compared with nymphalsurvival on water hyacinth. Of 50 nymphs placed onbanana, one developed to adulthood.
Of the native Pontederiaceae, feeding was recordedon Eichhornia natans but, compared with water hya-cinth, the plant produces very little emergent leafmaterial on which the nymphs can develop com-pletely. This species also has a slender petiole that issubmerged below the water and will not support ovi-positioning. Limited feeding and development wererecorded on Monochoria africana. The insects pre-ferred to feed on the epidermis of the petiole, andalthough this was damaging to the plant, it seemed toprovide the nymphs with insufficient nutrition todevelop. Full nymphal development was recorded onHeteranthera callifolia and although it was lower thanon water hyacinth it is still reason for concern.
Adult no-choice trials
Among the 16 species tested, oviposition wasrecorded on water hyacinth, M. africana and pickerelweed (Table 3). Only a few eggs were recorded on pick-erelweed, and only one eggpacket was recorded on M.africana. Oviposition probes are holes made by femaleslooking to lay eggs at the base of the petiole. Probeswere recorded on water hyacinth, M. africana and pick-erelweed. It appears as if the internal structure of the Mafricana petioles is not suitable for oviposition. Inseveral replicates of non-target species, egg cases werelaid on the sides of the cages and pots, indicating that thefemales were under oviposition stress and plants pre-sented to them did not offer suitable oviposition sites.
84
Table 1. Results of the first instar nympha host-specificity tests of Cornops aquaticum on selected plant species
Plant species No. Common name Feeding Development
AponogetonaceaeAponogeton distachyos L. 10 Cape pondweed 0 0
AlismataceaeAlisma plantago-aquatica L. 6 Water alisma 0 0
PoaceaeZea mays L.Arundo donax L.Phragmites australis (Cav.) Steud.Oryza sativa L.Saccharum officianum L.
101010
85
MaizeSpanish reedReedRiceSugarcane
000+0
00000
AraceaeZanthedeschia aethiopica (L.) Spreng.Colocasia esculenta L. SchottZamioculcas zamiifolia (Lodd.) Engl.Stylochiton sp.
2015
77
Arum LilyTaro
0000
0000
RestionaceaeElegia racemosa (Poir) Pers. 5 Restio 0 0
EriocaulaceaeEriocaulon dregei Hochst var sonderanium (Körn) Oberm. 5 0 0
CommelinaceaeCommelina africana L.Murdannia simplex (Vahl) Brenan
143
++
0+
PontederiaceaeEichhornia crassipes (Mart.) Solms-Laub.Eichhornia natans (P. Beauv.) Monochoria africana (Solms- Laub.) N.E.BrHeteranthera callifolia KunthPontederia cordata L.
45655
10
Water hyacinth
Pickerelweed
+++++
+0+++
JuncaceaeJuncus kraussi Hochst. subsp. krausii 5 Rush 0 0
ColchicaceaeGloriosa superba L. 7 Flame lily 0 0
AsphodelaceaeChlorophytum comosum (Thunb.) Jacq. 6 Hen and chickens 0 0
AlliaceaeAgapanthus africana (L.) HoffingAllium ampeloprasum (L.)Allium cepa L.
1055
AgapanthusLeekOnion
000
000
Continued on next page
85
Table 1. (Cont’d) Results of the first instar nympha host-specificity tests of Cornops aquaticum on selected plantspecies
Plant species No. Common name Feeding Development
LiliaceaeKniphofia linearifolia Bak.Tulbachia sp.Euricomis sp.Lillium sp.Bulbine sp.Aloe sp.Behnia reticulata DidrichsAsparagus officinalis L
6101010
6555
Red-hot poker 00000000
00000000
AmaryllidaceaeCrinum bulbispermum (Burm. f.)Clivia minata (Lindl.)Nerine sp.
1010
5
Orange River lilyBush lily
00+
00+
HypoxidaceaeHypoxis sp. 5 0 0
IridaceaeWatsonia sp. 5 0 0
MusaceaeMusa paradisica L. 10 Banana + +
CannaceaeCanna indica L.H. Bailey 10 Canna + +
ChenopodiaceaeBeta vulgaris L. var. cicla 10 Spinach 0 0
EuphorbiaceaeManihot esculenta Crantz 5 Cassava 0 0
BrassicaceaeRaphanus sativus L.Brassica oleracea L.Brassica rapa L.
1075
RadishCabbageTurnip
++0
+00
LeguminaceaePisum sativum L.Phaseolus vulgaris L.
1010
PeaBean
00
00
OnagraceaeLudwigia stolonifera (Guill. & Perr.) Raven 5 0 0
TrapaceaeTrapa natans L. var bispinosa (Roxb) Makino 5 Water chestnut 0 0
HalorgidaceaeLaurembergia sp. 5 0 0
Continued on next page
86
Field observations in the region of origin
Observations of host range were made at severallocalities in northern Argentina and in Peru.
In Argentina, 28 sites were surveyed. Of all theinsect species surveyed, at all the sites Cornops aquat-icum was considered to be the most damaging to waterhyacinth. Cornops aquaticum was also found to bewidespread and abundant on water hyacinth. Egg
cases were recorded on water hyacinth, Eichhorniaazurea and pickerel weed. The insect was found to beless abundant on pickerel weed, suggesting it is aninferior host. Cornops aquaticum was not recorded onCanna glauca or the two Commelina species evenwhen growing close to water hyacinth supporting highpopulations of the grasshopper.
In Peru, 30 sites were surveyed. Cornops aquaticumwas recorded on water hyacinth and Pontederia rotun-difolia. The grasshopper was abundant on P. rotundi-folia and caused severe damage to plants. Thepredaceous weevil, Ludovix fasciatus, was also found,and even with its presence Cornops aquaticum wasstill abundant.
Discussion
Cornops aquaticum is a very damaging natural enemyof water hyacinth and is likely to make a valuable con-tribution to the control of this weed in South Africa.This is evident from the fact that, despite being heavilyparasitised by the weevil Ludovix fasciatus in itsregion of origin, it is still abundant and damaging towater hyacinth. This weevil is not present in SouthAfrica, so it is predicted that the impact of the grass-hopper on water hyacinth would be greater.
Table 1. (Cont’d) Results of the first instar nympha host-specificity tests of Cornops aquaticum on selected plantspecies
a. Five first instar nymphs per replicate
Plant species No. Common name Feeding Development
ApiaceaeDaucus carota L. var. sativusHydrocotyle sp.
105
Carrot 00
00
SolanaceaeLycopersicon lycopersicum (L.)Solanum melogena L. var. sativusCapsicum annuum L.
101010
TomatoEggplantPepper
000
000
RubiaceaeCoffea sp. 5 Coffee 0 0
Cucurbitaceae Cucurbita pepo L.Cucumis sativus L.Citrillus lanatus (Thunb.)
555
MarrowCucumberWatermelon
000
000
AsteraceaeLactuca sativa L. var. capitata 10 Lettuce 0 0
Table 2. Mean number of Cornops aquaticum adultsreared from plant species during no-choicenymphal starvation trials
Plant species No. Mean number of adults/replicatea,b
a. Five first-instar nymphs were used per replicate.b. Figures in parentheses represent the standard deviation.
Eichhornia crassipes 45 3.47 (0.93)
Heteranthera callifolia 6 2.8 (1.21)
Pontederia cordata 10 1.60 (1.08)
Canna indica 10 1.10 (1.45)
Musa paradisica 10 0.02 (0.14)
87
The indigenous Eichhornia species in Africa, Eich-hornia natans, supports development of the grass-hopper nymphs, but the lack of emergent leaf materialand the submerged petioles suggest that the plant willnot sustain a population of C. aquaticum in the field.Of the other plants in the Pontederiaceae in Africa, M.africana does not support full development of thenymphs, and H. callifolia, although heavily attacked,did not support oviposition and is considered to beinferior to water hyacinth as a host.
Cornops aquaticum is considered to be oligopha-gous on Pontederiaceae and should be released only incountries that do not have native Pontederiaceae orwhere the spillover feeding on native Pontederiaceaewould be tolerable. Silveira Guido and Perkins (1975)found that, under high population levels in the labora-tory, nymphs fed on members of the Commelinaceae,rice and sugarcane in the Gramineae, and E. azureaand P. cordata in the Pontederiaceae. However, devel-opment was recorded only on Commelina spp. outsideof the Pontederiaceae. Under performance, or choicetests, they found that damage occurred to the sameCommelina species and to rice and sugarcane. Whilewe recorded some nibbling on rice, we have not
recorded any feeding on sugarcane. Furthermore,Bennett (1970) found that only water hyacinth wasattacked during choice tests with other species.
The host-specificity testing of this insect is incom-plete. However, despite relying on the most conserva-tive host-specificity tests (nymphal starvation trials)the insect has shown a high degree of specificity towater hyacinth.
Future Research
The emphasis in future research will be on the testingof the insect under more natural conditions. Thesetests might give less ambiguous results that wouldclarify the host specificity of C. aquaticum. Openfield trials in the region of origin are an option, whilewe believe that choice trials with adults and nymphswill clarify these results (Marohasy 1998). Tests willbe conducted using native Pontederiaceae from thesouthern African region, water hyacinth, canna andbanana. All these plants showed development of thenymphs. Special attention will be given to develop-ment and ovipositioning of Cornops aquaticum underopen field conditions.
Table 3. Mean number of Cornops aquaticum adults surviving and egg cases laid on test plant species during adult,no-choice trails. Two pairs of adults were used per replicate and each replicate lasted seven days.
Plant species Common name n Mean number of egg cases/ replicatea
a. Figures in parentheses represent the standard error.
Mean number of probes/ replicatea
Eichhornia crassipes Water hyacinth 8 5.21 (3.56) 3.67 (3.27)
Monochoria africana 6 0.04 (0.19 2.83 (1.72)
Heteranthera callifolia 6 0.00 ( – ) 0.00 ( – )
Eichhornia natans 3 0.00 ( – ) 0.00 ( – )
Pontederia cordata Pickerel weed 4 2.02 (0.80) 3.45 (1.67)
Canna indica Canna 8 0.00 ( – ) 0.00 ( – )
Musa paradisica Banana 6 0.00 ( – ) 0.00 ( – )
Commelina africana 8 0.00 ( – ) 0.00 ( – )
Murdannia simplex 4 0.00 ( – ) 0.00 ( – )
Zea mays Maize 3 0.00 ( – ) 0.00 ( – )
Raphanus sativus Radish 3 0.00 ( – ) 0.00 ( – )
Brassica oleracea Cabbage 5 0.00 ( – ) 0.00 ( – )
Nerine sp. 4 0.00 ( – ) 0.00 ( – )
Oryza sativa Rice 6 0.00 ( – ) 0.00 ( – )
Zanthedeschia aethiopica Arum lily 3 0.00 ( – ) 0.00 ( – )
Colocasia esculenta Taro 3 0.00 ( – ) 0.00 ( – )
88
Acknowledgments
We thank the Working for Water Programme of theDepartment of Water Affairs and Forestry for fundingof the project, and Mrs Isabel Roesch for collectingand propagating plants.
References
Bennett, F.D. 1970. Insects attacking water hyacinth in theWest Indies, British Honduras and the USA. WaterHyacinth Control Journal, 8, 10–13.
Center, T.D. 1994. Biological control of weeds: waterhyacinth and water lettuce. In: Rosen, D., Bennett, F.D.and Capinera, J.L., ed., Pest management in the tropics—a Florida perspective. UK, Intercept Ltd, 481–521.
Hill, M.P. and Cilliers, C.J. 1999. A review of the arthropodnatural enemies, and factors that influence their efficacy,in the biological control of water hyacinth, Eichhorniacrassipes (Mart.) Solms-Laub. (Pontederiaceae), in South
Africa. In: Olckers, T. and Hill, M.P. ed., Biologicalcontrol of weeds in South Africa (1990–1998). AfricanEntomology Memoir, 1, 103–112.
Jacot Guillarmod, A. 1979. Water weeds in southern Africa.Aquatic Botany, 6, 377–391.
Marohasay, J. 1998. The design and interpretation of host-specificity tests for weed biological control withparticular reference to insect behaviour. Biocontrol Newsand Information, 19, 13N–20N.
Perkins, B.D. 1974. Arthropods that stress water hyacinth.PANS, 20, 304–314
Silveira Guido, A. and Perkins, B.D. 1975 Biology and hostspecificity of Cornops aquaticum (Bruner) (Orthoptera:Acrididae), a potential biological control agent for waterhyacinth. Environmental Entomology, 4, 400–404.
Wright, A.D. and Purcell, M.F. 1995. Eichhornia crassipes(Mart.) Solms-Laubach. In: Groves, R.H., Shepherd,R.C.H. and Richerson, R.G., ed., The biology ofAustralian weeds. Melbourne, Australia, R.G. & F.J.Richerson, 111–121.
89
Establishment, Spread and Impact of Neochetina spp. on Water Hyacinth in Lake Victoria, Kenya
G.S. Ochiel*, S.W. Njoka*, A.M. Mailu† and W. Gitonga‡
Abstract
The Kenya Agricultural Research Institute imported 12,300 curculionid weevils (Neochetina spp.) fromdiverse sources, for biological control of water hyacinth in Lake Victoria, as part of the World Bank-fundedLake Victoria Environmental Management Project in East Africa. In addition to the rearing and quarantinefacility at Muguga, a second rearing facility was established in 1996 at Kibos, near Lake Victoria. The Kibosrearing facility and two community rearing facilities at the lakeshores, have produced approximately 100,000adult weevils and 42,000 weevil eggs over a three-year period. Since January 1997, some 73,500 Neochetinaweevils have been released at 29 sites and an additional 10,000 redistributed at several sites. Visualobservations and regular sampling monitored the establishment and spread and also evaluated the impact ofNeochetina weevils on water hyacinth. Within two years, weevils were established at 55% of release sites andwere being recovered 50 km from release sites. Post-release sampling data from four release sites in Berkeley,Kisumu and Kendu bays, indicated a reduction in leaf length, laminar area and fresh weight of water hyacinth,and a significant increase in number of weevil feeding scars and adult weevils per square metre. Three yearsafter the initial weevil releases, the combined mean number of weevils per plant for Kisumu, Nyakach, Kenduand Homa bays, was estimated to be six, well above the critical threshold of five weevils per plant. N. bruchiwas the dominant species accounting for 73.3% of the total weevil population. Thus, under Lake Victoriaconditions, the critical threshold was attained within 2–3 years of the initial releases.
LAKE Victoria (area ca 69,000 km2), shared by thethree East African countries, Kenya (6%), Uganda(43%) and Tanzania (51%), is the world’s second-largest freshwater lake (Figure 1). In 1989, it wasinvaded by water hyacinth and its presence in theKenyan part was confirmed in 1992. The origin of theinfestation is presumed to be in the River Kagera Basin
in Rwanda. At peak infestation in 1997, the areacovered by the weed in East Africa was more than15,000 ha. The tropical aquatic weed of South Amer-ican origin, has adverse impacts on the health, energy,water and transport sectors (Harley 1990; Harley et al.1996). The weed presented an enormous challenge forbiological control in East Africa.
As early as 1993, the Kenya Agricultural ResearchInstitute (KARI) imported water hyacinth weevils,Neochetina bruchi and N. eichhorniae, from the PlantHealth Management Division of the InternationalInstitute for Tropical Agriculture in Benin. These wee-vils, considered the most important biological controlagents against the water hyacinth, have had notablesuccess outside East Africa (Harley 1990; Julien andGriffiths 1998; Julien et al. 1999). However, host-spe-
* Kenya Agricultural Research Institute, National FibreResearch Centre Kibos, PO Box 1490, Kisumu, Kenya.Email: [email protected],[email protected]
† Kenya Agricultural Research Institute, PO Box 57811,Nairobi, Kenya. Email: [email protected]
‡ Kenya Agricultural Research Institute, NationalAgricultural Research Centre, Muguga, PO Box 30148,Nairobi, Kenya. Email: [email protected]
90
cificity tests were ordered in Kenya and Uganda,before releases in Lake Victoria were allowed.
Neochetina weevils were released in Lake Kyoga,Uganda in 1993 (Ogwang and Molo 1997, 1999) andin Lake Victoria in 1996 (James Ogwang, pers.comm.). The first weevil releases in Kenya were inLake Naivasha, which had water hyacinth since themid 1980s (Aggrey Mambiri, pers. comm.). Neo-chetina weevils were released in the Kenyan part ofLake Victoria in 1997 (Ochiel et al. 1999; Mailu et al.1999), while in Tanzania, Mallya (1999) reported thereleases of Neochetina weevils in the Pangani and Sigirivers in 1995, and in Lake Victoria in 1996.
This paper presents recent results from a program ofclassical biological control against water hyacinth inLake Victoria, implemented by KARI under the LakeVictoria Environmental Management Project.
Materials and Methods
Mass rearing and releases of Neochetina spp.
Since 1996, Kenya Plant Health and InspectorateServices has allowed KARI to import adult N. bruchiand N. eichhorniae from Uganda, South Africa andAustralia for the biological control of water hyacinthin Lake Victoria. KARI established a second weevilrearing facility in December 1996, at the NationalFibre Research Centre (NFRC), Kibos, near LakeVictoria. ‘Breeding stock’ for the Kibos rearingfacility was obtained from the quarantined mass-rearing facility at the National Agricultural ResearchCentre, Muguga, near Nairobi. The breeding materialconsisted of mature adult Neochetina weevils andhost plants inoculated with weevil eggs. Later, adult
0 40 80 120 160 200 kilometres
Bukoba
Kigali
Mwanza
Homa
Kisumu
Kyaka
RWANDA
BURUNDI
KENYA
TANZANIA
UGANDA
LAKE VICTORIAKagera R.
Victo
riaN
ile
Mig
ori R
.
KageraNational
Park
Lake Kyoga
LakeNaivasha
Figure 1. Lake Victoria and surrounds
91
Neochetina weevils were imported from Uganda formass rearing. Julien et al. (1999) describe in detailrearing and harvesting techniques for Neochetinaweevils from plastic tubs, rearing pools and galva-nised corrugated iron sheet tanks, all of which havebeen in use at the Kibos rearing facility. Addition-ally, ‘Technotank’ PVC tanks (120 × 60 cm; 230 L),with sawn-off lids, have been used to rear the weevilsat Sango Rota and Nyamware beaches and at OgenyaPrimary School (community-based rearing facilitiesnear the lake). Fertiliser NPK 17:17:17 and driedcow-dung were added to the rearing containers oncea month to maintain plant vigour.
Weevils were harvested for field releases asdescribed by Julien et al. (1999). Neochetina weevilsimported from South Africa were released in LakeVictoria in 1997 and further releases were carried outwith weevils reared at NFRC Kibos and communityrearing facilities. Hyacinth plants infested with weevillife stages and adult weevils were used for releases.Adult weevils were fed on fresh leaves and petioles inplastic jars before transporting them to release sites.Release techniques included planting host plantsinfested with weevil life stages among hyacinth plantsand tipping adult weevils from the plastic containersonto hyacinth plants. Weevils were also released atsites more than 50 m from the shoreline. Canoes wereused to release at sites that were inaccessible by motorvehicle or on foot.
Monitoring the establishment and spread of Neochetina weevils
We recorded petiole damage by weevil larvae,fresh adult feeding scars and the number of adultweevils on water hyacinth at release sites with resi-dent mats of water hyacinth and at non-release sites.
These visible signs are indicators of an establishing orestablished weevil population at a given site. Weevilrecovery at non-release sites indicated weevil spreadon water hyacinth.
Evaluation of the impact of Neochetina spp. weevils on water hyacinth
Using a modified sampling protocol developed atthe Commonwealth Scientific and IndustrialResearch Organization (CSIRO), Australia, we eval-uated the impact of Neochetina spp. on water hya-cinth. The objectives of the sampling were to: (1)evaluate water hyacinth growth parameters; (2)quantify weevil feeding damage; and (3) estimateweevil populations. A half-metre2 quadrat wasthrown randomly on mats of hyacinth plants. Thenumber of plants per quadrat was recorded. For eachof 10 or 30 plants from the quadrat or nearby, the fol-lowing parameters were recorded: fresh weight; leaflaminar area; leaf length; number of feeding scars;number of weevils per plant; and number of adultweevils per square metre (mean number of weevilsper plant × number of plants per quadrat). Rapidassessment of weevil populations was done bycounting the number of weevils from each of 10 or 20randomly selected plants at selected sites.
Results
Importation of Neochetina weevils
Between 1996 and 1998, KARI imported 12,300Neochetina weevils from Australia, South Africa andUganda, for mass rearing and releases on water hya-cinth in Lake Victoria (Table 1).
Table 1. Importations into Kenya of Neochetina weevils for biological control of water hyacinth in Lake Victoria
Species Year imported Number Purpose Source
Neochetina bruchi 1996 1300 Mass rearing Uganda
1997 2000 Mass rearing/releases Australia
1998 1000a
aBatch did not survive
Releases South Africa
Neochetina eichhorniae 1997 5000 Mass rearing/releases South Africa
1997 2000 Mass rearing/releases Australia
1998 1000 Releases South Africa
Total 12,300
92
From December 1996 to December 1999, the Kibosrearing facility and community rearing facilities pro-duced approximately 100,000 adult weevils, of which25,000 were for ‘breeding stock’ and for releases inLake Naivasha and Nairobi Dam.
Between January 1997 and December 1999,approximately 73,200 adult weevils were released at29 sites in Kisumu, Nyando, Rachuonyo, Bondo,Homa Bay, Migori, Suba and Busia districts (Table 2).An additional 10,000 weevils were redistributed fromthe Homa Bay Pier and Police Pier release sites toother sites within the Kisumu District.
Monitoring the establishment and spread of Neochetina weevils
Monitoring in January 1999 confirmed that theweevils were firmly established at 16 sites in 7 dis-tricts along the Lake Victoria shoreline (equivalent to55% of the sites). Weevil recoveries were also madeat distances ranging from 5–50 km from the nearestrelease sites.
Evaluation of the impact of Neochetina spp. weevils on water hyacinth
In general, post-release sampling data collected(November 1997 to May 1998) at four selected releasesites in Berkeley, Kisumu and Kendu Bays, indicateda suppression of plant growth parameters (freshweight, leaf laminar area and leaf length) and substan-tial increases in number of feeding scars and adultweevils per plant (Table 3). Fresh weight reductionwas noted at a single site, Bukoma Beach. Leaf lengthreduction was noted at two sites, while leaf laminararea reduction was evident at Sio Port and Bukoma.The number of feeding scars and adult weevils perplant increased at all sites.
Estimations of weevil populations
Post-release sampling of water hyacinth at sixselected sites in three bays (May–December 1999),gave a combined mean number of 6.0 Neochetinaweevils per plant, with actual number of weevils perplant ranging from 0 to 32 (Table 4). Table 4 alsoshows that N. bruchi was the dominant of the twoweevil species, accounting for 73.3% of the totalweevil population.
Discussion
Importation of additional biological control agents, themoth Niphograpta albiguttalis, the mite Orthoga-lumna terebrantis and the hemipteran bug Eccrito-tarsus catarinensis, to augment biological controlefforts by Neochetina weevils, is recommended.Rearing pools, which are easier to manage and have alarger capacity, are preferred over both plastic basinsand tanks and galvanised iron sheet tanks. Tub rearingwas found to be labour-intensive and time-consuming.Tubs may, however, be used for ‘demonstration massrearing units’ in schools and community-based rearingfacilities near the lake.
Releases on floating mats assisted in the redistribu-tion and spread to non-release sites. Wind and watercurrents were responsible for the spread of weevils onfloating mats of water hyacinth. Under the environ-mental conditions of Lake Victoria, weevils estab-lished quite rapidly.
At a regional level, monitoring the water hyacinthinfestation pattern using aerial reconnaissance photog-raphy, ground truthing and satellite imagery has beenproposed. At a national level, monitoring and evalua-tion of the impact of weevils on water hyacinth, redis-tribution to areas with low weevil populations andscouting for new infestations should continue.
Weevil damage has been held primarily responsiblefor the reduction of the weed cover by up to 80%, fromthe peak infestation of 6000 ha in 1998 (Synoptics,Integrated Remote Sensing and GIS Applications, TheNetherlands). By late 1999, water hyacinth in theKenyan part of the lake was no longer capable of flow-ering and producing ramets (daughter plants). This hasbeen attributed to weevil damage and opportunisticfungi. The El Niño flooding of 1997 may have physi-cally destroyed plants by washing them ashore.
Ecological succession of water hyacinth by emer-gent plant species, mainly papyrus (Cyperus papyrus)and hippograss (Vossia cuspidata), has been noted(Ochiel, personal observations). This phenomenon hasalso been observed in Lake Kyoga, Uganda, followingthe successful biological control of water hyacinth byNeochetina weevils. However, this is short-lived andthe secondary vegetation will disappear after thedegraded hyacinth substratum supporting it eventuallysinks.
The long-term approach to water hyacinth manage-ment and indeed other floating or submerged aquaticweeds, should focus on curbing the discharge of efflu-ents into Lake Victoria from surrounding urban settle-ments, agricultural and industrial activities.
93
Table 2. Releases of Neochetina at sites in Lake Victoria, Kenya, January 1997 to September 1999.
Site–Grid Reference Release dates Life stage
Eggs Adults
Police Pier 0°5.5'S;34°44.3'E 23.1.97–18.2.98 13 850 5 553
Fisheries Pond 0°5.4'S;34°44.0'E 23.1.97–15.4.97 3 680 1 000
Golf Club 0°5.4'S;34°43'E 23.1.97–18.5.97 600 500
Yacht Club 0°8.5'S; 34° 45.5'E 22.1.97–3.6.98 3 500 6 705
Usoma Beach 0°06'S;34°38'E 21.2.97–14.7.98 10 695 5 967
Karamadhan 0°07'S;34°38'E 27.2.97–15.4.98 6 250 750
Otonglo Beach 0°04S;34°39.5'E 21.7.97–7.6.99 3 100 3 071
Dunga Beach 0°09'S;34°46.5'E 23.5.98–6.8.98 3 680 2 042
Kaloka Beach 0°9.5'S;34°32.5'E 17.5.98–28.5.98 2 075
Sango-Rota Beach 0°16.5'S, 34°47.5'E 7.6.97–25.3.98 2 153
30.7.99–30.9.99 10 000
Kusa Beach 0°18.5'S,34°51'E 17.11.98 1 066
Nduru Beach 0°15.5'S;34°51.5'E 21.5.98 979
Kendu Bay Pier 0°20'S;34°39'E 7.6.97–10.8.98 2 150
K’Owuor Pier 0°21'S;34°28'E 30.1.98 740
Homa Bay Pier 0°31'S;34°28'E 21.11.97 509
Ombogo Beach 0°28.5'S;34°30'E 29.1.98 430
Tagache Beach 0°58'S,34°6.5'E 28.1.98 100
Sori-Karungu Beach 0°50'S,34°10'E 29.1.98 300
Luanda Nyamasare Beach 0°27'S,34°17'E 21.4.98 508
Aram Beach 0°18'S,34°16'E 12.11.97–11.3.98 1 250
Usenge 0°03'S,34°05'E 12.11.97–16.5.98 1 250
Usigu (Uharia) Beach 0°04'S,34°9.5'E 3.2.98–16.5.98 1 750
Obenge Beach 0°13'S,34°12.5'E 16.5.98 540
Luanda Kotieno 0°18'S,34°16'E 11.3.98 250
Sio Port 0°14'N,34°02'E 12.9.97–13.1.99 1 200
Bukoma Beach 0°12'N,33°58.5'E 29.9.97–2.2.98 1 046
Nyamware Beach 0°16'S, 34°42'E 13.8.98 950
12.7.99–30.9.99 15 000
Ogenya Beach 0°15.5'S,34°52'E 20.5.99–22.5.99 471
Total 41 975 73 225
94
Table 3. Post-release sampling data to evaluate the impact of Neochetina weevils on water hyacinth at four sites inLake Victoria, Kenya
Site Sampling date Fresh weight (g) ± SE
Leaf length(cm) ± SE
Laminar areaa (cm2) ± SE
aSecond youngest petiole sampled
Feeding scars ± SE
Weevils/plant ± SE
Sio Port (40 m2)
19.11.97 1685±958 137.2 ± 14.9 195.4±9.7 2.5 ± 1.9 0.4 ± 1.4
10.3.98 3550±1755 77.8 ± 23.0 110.2±12.5 100.3 ± 9.6 1.8 ± 2.6
Bukoma Beach (15 m2)
20.11.97 2270±935 162.9 ± 16.5 178.6±0.7 2.5 ± 2.1 0.2 ± 0.6
10.3.98 925±528 75.5 ± 19.9 126.8±13.0 107.4 ± 28.2 2.2 ± 1.9
Police Pier (1500 m2)
28.11.97 251±128 19.9 ± 4.8 49.0±5.0 19.4 ± 6.9 0.4 ± 0.5
20.5.98b
bn = 30. At all other sites n = 10.
482±271 31.3 ± 15.8 74.6±12.8 138.8 ± 28.3 4.5 ± 3.9
Kendu Bay Pier (400 m2)
21.11.97 1950±797 78.8 ±19.3 146.8±5.6 2.9 ± 2.3 0.1 ± 0.3
12.3.98 2510±127 100.3 ±33.1 124.8±13.1 268.3 ± 52.4 6.0 ± 3.0
Table 4. Neochetina weevil populations on water hyacinth estimated from six sites in Lake Victoria, Kenya, May–December 1999.
Site Sampling date Mean no. of weevils/planta
± SE
aMean of 10 plants per site, except for Homa Bay 18.9.99, where n=20
Mean no. of weevils by speciesb ± SE
bNb = Neochetina bruchi. Ne = N. eichhorniae.
Range
Nb Ne
Kisumu Bay
Police Pier 6.5.99 2.5 ± 2.3 1.8 ± 0.9 0.7 ± 0.3 1–4
14.12.99 6.3 ± 4.5 5.1 ± 3.3 1.2 ± 0.6 0–6
Karamadhan 6.5.99 1.8 ± 2.1 1.1 ± 1.7 0.7 ± 0.9 0–6
4.12.99 3.7 ± 2.8 2.9 ± 2.3 0.8 ± 0.9 0–7
Nyakach Bay
Kusa 7.5.99 14.0 ± 6.7 14.0 ± 6.7 0.0 ± 0.0 2–22
14.12.99 3.2 ± 3.9 3.2 ± 3.9 0.0 ± 0.0 0–11
Sango Rota 7.5.99 5.4 ± 4.4 3.7 ± 2.9 1.7 ± 2.3 1–13
15.12.99 2.9 ± 1.6 1.5 ± 1.9 1.4 ± 1.6 0–8
Kendu Bay
Kendu Bay 7.5.99 2.4 ± 2.3 2.0 ± 1.7 0.4 ± 0.7 0–6
Pier 16.12.99 2.0 ± 1.7 1.5 ± 1.6 0.5 ± 0.9 0–4
Homa Bay
Homa Bay 18.9.99 18.1 ± 15.3 15.4 ± 7.2 2.7 ± 2.3 0–32
Pier 15.12.99 9.2 ± 8.6 6.5 ± 5.9 2.7 ± 3.0 0–32
Grand mean 6.0 ± 5.3 4.4 ± 3.1 1.6 ± 0.9
Percentage 73.3 26.7
95
Acknowledgments
The authors wish to acknowledge financial supportfrom the KARI, GTZ, FAO, AusAID and the WorldBank, through the Lake Victoria Environmental Man-agement Project. Technical advice from Mic Julienand Tony Wright, Commonwealth Scientific andIndustrial Research Organisation, Australia is grate-fully acknowledged. Technical assistance fromS. Akello, S. Lwili, A. Lumumba, D. Mutisya, J.Ogada and G. Omito is highly appreciated.
References
Harley K.L.S. 1990. The role of biological control in themanagement of water hyacinth. Biocontrol News andInformation, 11, 11–22
Harley, K.L.S., Julien, M.H. and Wright, A.D. 1996. Waterhyacinth: a tropical world-wide problem and methods forits control. Proceedings of the 2nd International WeedControl Congress, Copenhagen, 639–643.
Julien, M.H. and Griffiths M.W. 1998. Biological control ofweeds; a world catalogue of agents and their target weeds,4th ed. Wallingford, UK, CAB International, 223p.
Julien, M.H. and Griffiths, M.W. and Wright, A.D. 1999.Biological control of water hyacinth: the weevilsNeochetina bruchi and N. eichhorniae, biologies, hostranges and rearing, releasing and monitoring techniquesfor biological control of Eichhornia crassipes. Canberra,Australian Centre for International AgriculturalResearch, ACIAR Monograph No. 60, 87p.
Mailu, A.M., Ochiel, G.R.S., Gitonga, W. and Njoka, S.W.1999. Water hyacinth: an environmental disaster in theWinam Gulf and its control. In: Hill, M.P., Julien M.H.and Center, T.D., ed., Proceedings of the 1st WorkingGroup Meeting for the Biological and Integrated Controlof Water Hyacinth, 16–19 November 1998, Harare,Zimbabwe. Pretoria, South Africa, Plant ProtectionResearch Institute, 115–119.
Mallya, G. 1999. Water hyacinth control in Tanzania. In:Hill, M.P., Julien M.H. and Center, T.D., ed., Proceedingsof the 1st Working Group Meeting for the Biological andIntegrated Control of Water Hyacinth, 16–19 November1998, Harare, Zimbabwe. Pretoria, South Africa, PlantProtection Research Institute, 25–29.
Ochiel, G. R. S., Mailu, A. M., Gitonga W. and Njoka, S. W.1999. Biological control of water hyacinth on LakeVictoria, Kenya. In: Hill, M.P., Julien M.H. and Center,T.D., ed., Proceedings of the 1st Working Group Meetingfor the Biological and Integrated Control of WaterHyacinth, 16–19 November 1998, Harare, Zimbabwe.Pretoria, South Africa, Plant Protection ResearchInstitute, 101–105.
Ogwang, J.A. and Molo, R. 1997. Biological control ofwater hyacinth in Uganda. 6th East African BiennialWeed Science Conference Proceedings, 247–253.
Ogwang, J.A. and Molo, R. 1999. Impact studies onNeochetina bruchi and Neochetina eichhorniae in LakeKyoga, Uganda. In: Hill, M.P., Julien M.H. and Center,T.D., ed., Proceedings of the 1st Working Group Meetingfor the Biological and Integrated Control of WaterHyacinth, 16–19 November 1998, Harare, Zimbabwe.Pretoria, South Africa, Plant Protection ResearchInstitute, 10–13.
96
Water Hyacinth Population Dynamics
J.R. Wilson,* M. Rees†, N. Holst,‡ M.B. Thomas* and G. Hill§
Abstract
Neochetina eichhorniae and N. bruchi have in some locations been very successful in controlling water hyacinthinfestations. Understanding the conditions under which the weevils are not successful is a key area of research.We have used simple analytical tractable models to investigate this problem. We argue that biomass density andpercentage coverage are the two most useful parameters to measure in a monitoring program. We modelledwater hyacinth as a population of biomass. Under stable conditions, the logistic growth model accuratelydescribes water hyacinth growth. Understanding how abiotoic conditions alter the parameters of the model isessential for accurate prediction of water hyacinth growth. There appear to be five main factors limitinginfestations of water hyacinth: salinity, temperature, nutrients, disturbance and natural enemies. The models aremodified to include the effect of weevil damage. Simple deterministic models are developed that incorporatedevelopmental delays and population stage structure. For realistic parameter values, the models predicteradication of water hyacinth. We discuss how this prediction is altered in a dynamic environment. The factorsthat may limit the weevil population under stable conditions, and so prevent eradication, are explored. In orderto test these ideas, information on areas where control has and has not been successful needs to be collated.
THE current status of water hyacinth control has beenwell reviewed in these proceedings and elsewhere(Julien et al. 1996, 1999). Models have been used toinvestigate the effect of different management strate-gies (Ewel et al. 1975; Mitsch 1976; Lorber et al. 1984;Musil and Breen 1985b) drawing on the wealth ofinformation from many studies conducted worldwide.However, there has been only one published modelinvestigating the effect of biological control agents(Akbay et al. 1991). Models used to understand whenan insect biological control agent will control a weedhave produced insights into how control can beachieved (Lonsdale et al. 1995; Rees and Paynter1997). Furthermore, the Lotka–Volterra model hasbeen successfully used to simulate the growth ofanother aquatic weed (Salvinia molesta) before andduring control by Cyrtobagous salvinae (Room 1990).This approach can be useful in drawing together
existing knowledge, as well as to identify areas whereresearch needs to be concentrated.
The aim of the current research is to develop a pre-dictive model for the control of water hyacinthaddressing the following questions:
• what causes variability in water hyacinthinfestations?
• how does the introduction of Neochetinaeichhorniae affect the size of the infestation?
• what can be done to improve control?
In this paper we will outline some preliminaryresults of modelling work and the hypotheses gener-ated from this work.
Modelling Water Hyacinth Biomass
When building a model, an appropriate state variable,which describes the state of the system at any momentin time, must be selected. With animal populations, thestate variable chosen is usually the number of individ-uals. When modelling diseases, the number infected,
* NERC Centre for Population Biology, Imperial College atSilwood Park, Ascot, Berks, SL5 7PY, UK.Email: [email protected]
97
infectious, immune and susceptible can be used tocharacterise the host population. With plant popula-tions, the choice of state variable is often less clear. Inthis section, we explain why we have chosen biomassas the state variable.
The current model will be most useful if the statevariable reflects the magnitude of the problem. Waterhyacinth infestations have negative impacts on health,food production, navigation, hydroelectric schemes,irrigation schemes and recreation (see e.g. Gopal1987). These problems are caused by the sheer bulk ofvegetation and the fact that the vegetation coversgreat areas. As the scale of the problem depends onthe bulk of the weed, biomass would be an appropriatestate variable.
Most studies have been conducted using biomass orindividual density. The density of individuals is more
easily determined than biomass, but the point at whichan offshoot becomes a separate plant is not alwaysclear. The main disadvantage of using individuals isthe great variability in the size of an individual.Madsen (1993), using experimental data, proposed ahumped relationship between biomass density andindividual plant density. We have plotted these resultsand some from another study in Florida (Center andSpencer 1981) in Figure 1. Similar patterns are shownin several unpublished data sets (M. Purcell, unpub-lished data; T. D. Center, unpublished data). Plantsgrown at low density have relatively constant biomass.However, at densities above about 500 g (dry weight)per square metre there is no clear relationship betweenindividual density and biomass. This suggests thatbiomass provides a better description of the scale ofthe problem than individual plant density.
A
0
500
1000
1500
2000
2500
3000
3500
Nu
mb
er o
f pla
nts
/ m
2 N
um
ber
of p
lan
ts /
m2
B
0
50
100
150
200
250
300
350
0 500 1000 1500 2000 2500 3000
0 100 150 200 250 30050
Biomass density / g (dry weight) per m2
Biomass density / g (dry weight) per m2
Figure 1. Biomass density plotted against individual density: A (Center and Spencer 1981); B (Madsen 1993)
98
Importantly, studies have shown little variation inplant water content (average of 94–95%), althoughthere is some variation between studies and betweenthe different plant parts (Penfound and Earle 1948;Sahai and Sinha 1970; Debusk et al. 1981). Therefore,it is straightforward to convert between dry and freshweight.
It is possible that other units, e.g. petiole length, canalso be converted to biomass, and if these units areeasier to measure they might be a more appropriateunit for study. Center and Spencer (1981) found thatplant weight in a lake in Florida was closely related tothe mean number of leaves per plant and the meanmaximum leaf length (including the petiole). Itremains to be confirmed whether this relationshipholds when weevils are present. Moreover, at differentnutrient levels, the ratio of biomass in the roots andshoots is different, and so the proposed relationshipmay be different. However, finding a reliable surro-gate would allow easier monitoring and allow themodels to be tested using existing data-sets that do notcontain information on biomass.
One of the assumptions of the modelling work isthat populations of water hyacinth around the worldare not genetically different with respect to growth.Clonal differences have been investigated (Watsonand Cook 1987) and currently the genetic variationbetween populations of water hyacinth worldwide isbeing assessed as part of the IMPECCA project(Bateman, these proceedings). These studies suggestthere is some variation, especially with flowering, buta simple growth model based on biomass should begenerally applicable.
Logistic Model of Water Hyacinth
Here we discuss a simple model and how to parame-terise this model using data. Understanding how envi-ronmental conditions affect the parameters of themodel will be useful in predicting the size of waterhyacinth infestations. We have modelled the growth ofwater hyacinth using a logistic model, equation 1 (alsosee Gutiérrez et al. 2001).
The biomass density of plant material is P (g (dryweight)/m2) and dP/dt is the rate of change of the pop-ulation. This model has two parameters: the intrinsicgrowth rate, r, and the carrying capacity, K. At lowdensities the population will increase at its intrinsic
rate of growth, r. As the density of plants approachesthe carrying capacity, K, the rate of increase in the pop-ulation, dP/dt, tends linearly to zero. Furthermore, ifthe plant density is above the carrying capacity, thenthe population will fall to K. With constant parameters,this model has a stable point equilibrium at K and anunstable point equilibrium at zero, providing r > 0(May 1981).
Both r and K are estimated from field and laboratorystudies. Changes in biomass with time, r, have beenmeasured in many different situations. We have alsoestimated the carrying capacity using the highestlevels seen in nature and in long-term experiments. Inboth cases, we have expanded on the review of thewater hyacinth growth parameters reported by Gopal(1987). Of these studies, those that have been carriedout at several plant densities have been used to esti-mate both parameters. One of the assumptions of thelogistic model is a negative linear relationshipbetween plant density and intrinsic rate of growth. Formost situations this gives a reasonable fit (Fig. 2).However, there is some curvature (Fig. 2C) whichwould imply an under-estimation of K. Within a siteand season this model shows a good fit with experi-mental data. However, between studies and betweenseasons (Fig. 2C) there is variation in both themaximum intrinsic rate of growth r and the carryingcapacity K. This variation reflects the variation inwater hyacinth infestations and indirectly how waterhyacinth is affected by the environmental conditions.
How the Environment Affects the Parameters of the Logistic Model
The conclusions from the parameterisation of thelogistic model appear qualitatively similar to previousreviews. There appear to be five main factors limitingthe growth rate and carrying capacity of water hya-cinth: salinity, temperature, nutrients, disturbance andnatural enemies (in the host range of water hyacinth).
• Salinity—water hyacinth is killed in waters that aremore than about 0.2% saline (Haller et al. 1974:Nwankwo and Akinsoji 1988). This is important inestuarine areas e.g. the coastal lagoons of WestAfrica.
• Low temperature—stops the weed establishing intemperate areas and prevents it from reaching highlevels in the sub-tropics e.g. California (Bock1966). From their experimental study, Knipling etal. (1970) proposed a parabolic relationshipbetween temperature and growth rate, with growth
dP
dtr P
P
K= −
. . 1 (1)
99
tailing off quickly after the optimum of 30°C.Imaoka and Teranishi (1988) proposed that rincreases exponentially with ambient temperaturesin the range 14 to 29°C. Both models predict thatwater hyacinth growth stops below 13°C. However,field populations of water hyacinth may be morelimited by frost damage, as this increases the loss ofbiomass.
• Nutrients—the levels of available nitrogen andphosphorus have been often cited as the mostimportant factors in limiting water hyacinth growth(Carignan and Neiff 1994; Heard and Winterton2000; Musil and Breen 1985a; Reddy et al. 1989,1990, 1991). The half-saturation co-efficients forwater hyacinth grown under constant conditionshave been found to be from 0.05 to 1 mg/mL fortotal nitrogen and from 0.02 to 0.1 mg/mL for
phosphates. Water hyacinth growth quickly tails offbelow the lower limits. The effect of other mineraldeficiencies has also been studied (Newman andHaller 1988).
• Disturbance—flooding can break up large mats ofwater hyacinth and leave plants stranded on land.Similarly, currents flush water hyacinthdownstream. However, water hyacinth can stillbuild up on sheltered edges and at blockages. Waveaction may itself limit growth by directly damagingplants and by forcing the weed to maintainaerenchymous tissue.
• Natural enemies—in its native range in SouthAmerica, water hyacinth is controlled by a suite ofnatural enemies. It can be the dominant floatingaquatic weed but not always and not everywhere(H. Evans, pers. comm.).
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.10
0 500 1000 1500 2000
Spec
ific
gro
wth
rate
(g p
er g
per
day
)
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.10
Spec
ific
gro
wth
rate
(g p
er g
per
day
)
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.10
Spec
ific
gro
wth
rate
(g p
er g
per
day
)
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.10Sp
ecifi
c g
row
th ra
te (g
per
g p
er d
ay) C D
A B
0 500 1000 1500 2000
Density / g (dry weight) per m2
Density / g (dry weight) per m2
0 500 1000 1500 2000
0 500 1000 1500 2000
Density / g (dry weight) per m2
Density / g (dry weight) per m2
Jan–Feb R2 = 0.88
Sept–Mar R2 = 0.88
Aug–July R2 = 0.90
y = –0.000074x + 0.068422
R2 = 0.694275
y = –0.000038x + 0.083037
R2 = 0.958085
y = –0.000040x + 0.081525
R2 = 0.995590
Figure 2. r against K measured experimentally in: A, Japan (Imaoka 1988); B, Florida, USA (Reddy 1984); C, Argentina (Fitzsimons 1986); and D, Florida, USA (Debusk 1981)
100
Modelling Water Hyacinth and N. eichhorniae
We now modify the model to investigate the introduc-tion of a weevil biological control agent, Neochetinaeichhorniae, to water hyacinth in its exotic range. Theweevils do not have discrete generations, althoughwinter or a severe event may synchronise a population.Moreover, plant biomass production is a continualprocess. Therefore, a continuous time model was used.Caughley and Lawton (1981) presented a range ofplant/herbivore models e.g. equation 2.
The equation for the plant population is the logisticgrowth model with a loss term due to weevil feeding:c1.A.(1 – e–d1.P). The weevil population, A, increasesat the maximum rate –a + c2 when there are manyplants (i.e. e–d2.P is approximately zero), and declinesat the maximum rate of a when there are few plants.Using parameters from the literature (Center andDurden 1986; Jayanth and Visalakshy 1990; Heardand Winterton 2000), this model predicts that waterhyacinth will very quickly be eradicated (Fig. 3A).
This model assumes all weevils have the same effecton the plant. However, late larval stages are the mostdamaging. To mimic this we have added a time delayto the growth of the weevil population. Under theseconditions, the weevil no longer drives water hyacinthto extinction, but instead the system undergoes largeamplitude cycles. Water hyacinth is driven toextremely low densities during these cycles, which
dP
dtr P
P
Kc A e
dA
dtA a c e
d P
d P
= −
− −( )= − + −( )( )
−
−
. . . .
. .
.
.
1 1
1
1
2
1
2
(2)
A
0
500
1000
1500
2000
2500
3000
3500
0
500
1000
1500
2000
2500
3000
3500
0
500
1000
1500
2000
2500
No
. of w
eevi
ls
B
0
100,000
200,000
300,000
400,000
500,000
600,000
No
. of w
eevi
ls
PlantsHerbivore
PlantsHerbivore
0 10 20 30 40 50 60 70
Days
0 100 200 300 400 500 600
Days
Plan
t b
iom
ass
(g(d
ry w
eig
ht)
per
m2 )
Plan
t b
iom
ass
(g(d
ry w
eig
ht)
per
m2 )
Figure 3. Model outputs for models with A, no time delay, and B, with a time delay
101
would effectively result in extinction. This simplemodel assumes the weevil population can be charac-terised by the number of larvae. To improve therealism of the model we have added stage structure.The weevil population is approximated by ‘… asequence of developmental stages within each ofwhich all individuals can sensibly be regarded as func-tionally identical (that is all having the same per capitavital rates)’ (Gurney et al. 1983). This makes themodels more difficult to analyse, but again the weevilsappear to eradicate the plant.
These models predict that, given stable conditions,water hyacinth will always be controlled. However,from field sites this is known not to be the case(T. Center, pers. comm.). Therefore, the modelsappear to overemphasise the effect of the weevils andso in some way fail to capture an important aspect ofthe water hyacinth/weevil interaction.
What Limits the Weevils?
In this section, we investigate why control is notalways as predicted. We first discuss how a fluctuatingenvironment can, within the framework of the existingmodels, prevent water hyacinth from being eradicated.Then we move onto what happens in stable situationswhere the models do not give the correct qualitativeconclusion.
Under certain dynamic scenarios, the weevils mayhave little impact on the water hyacinth population.Frost kills leaves, which in turn kills weevil eggs andyoung larvae. However, late larvae, pre-pupae andpupae may survive around the rootstock. When plantsbegin to regrow in the spring, the weevils need tofinish maturing, mate, oviposit, and develop beforethe next late larval stages can cause major damage tothe plants. This developmental delay in the weevilpopulation may allow the plant to outgrow weevildamage, providing the plant has not been too heavilydamaged by the frost. Factors that speed the waterhyacinth regrowth, e.g. high nutrients, may exacer-bate this. Alternatively, water hyacinth grows at tem-peratures lower than the weevils and so the weed isfreed from herbivore pressure early in the growingseason. The effect of cold on the weevil populations iscurrently under investigation (M. Hill, pers. comm.).Herbicidal or mechanical control may cause similarproblems by removing the age structure in the weevilpopulation and, for a short time, freeing the weedfrom herbivore pressure. Extreme natural conditionsmay also have an adverse effect on control. Floodingcan bring new plants from upstream or remove
weevil-infested plants from a population. Droughtmay dramatically reduce the population of plants andweevils. After drought, water hyacinth seeds will ger-minate with rising water levels and these new weevil-free plants can re-establish the problem before theweevil population can respond (Guillarmod andAllanson 1978). Furthermore, an infestation maypersist despite a high weevil population if new plantscontinually arrive from upstream.
The lack of control under stable conditions, how-ever, suggests there is something limiting the weevils.Here we explore where density dependence may beacting in the weevil population—at the egg, the larval,the pupal or the adult stage. Oviposition sites are prob-ably not limiting and at very high egg densities the fer-tility of the eggs does not appear to change.Furthermore, there are few records of egg parasitismor predation, and so it is unlikely to occur at the eggstage. Larval cannibalism has been used as a possibleexplanation (DeLoach and Cordo 1976), but the fewaccounts of this suggest it is a rare, accidental phenom-enon caused by larger larvae accidentally tunnelingthrough smaller larvae (T. Center and M. Julien, pers.comm.). Larval competition for food may directlyincrease larval mortality; indirectly increase mortalityby prolonging larval duration; or reduce the size oflarvae at pupation and thereby increase pupal mor-tality or decrease adult fecundity (Gurney and Nisbet(1985) presented some models illustrating these).Larval damage can result in the flooding and sheddingof a petiole, which would kill any larvae remaining init (Center 1987). This sort of asymmetrical competi-tion would occur only at high damaging densities, butshould result in a few of a given age group developing.In order to elucidate this, we have undertaken anexperiment to measure the effect of egg density onlarval development at two different water nutrientlevels.
Pupae, or pre-pupae, may be the limiting stagebecause, even accounting for the shorter stage dura-tion, pupal cocoons in the field are often less commonthan larvae (M. Julien, pers. comm.). Pupal mortalitymay be higher in silted water or where the plant rootsused by the pre-pupae are buried in the sediment.Muddy edges to a water-body appear to be correlatedwith unsuccessful control (Visalakshy and Jayanth1996; O. Ajounu, pers. comm.). Experimentally,larvae have been shown to develop and cause damageto rooted plants (Forno 1981). Pupae have also beenfound on rooted plants in the field (M. Hill, pers.comm.). However, Visalakshy and Jayanth (1996)found that larvae on plants with silted roots were a
102
third as likely to survive to adults as larvae on freefloating plants, although once pupae had formed acocoon there was no difference in survival. They pro-posed that, in silted conditions, there was either ashortage of pupation sites or the pre-pupae have amuch lower success in forming cocoons on siltedroots.
Finally, the limitation may occur at the adult stagethrough either emigration or mortality. Adults havebeen observed to develop flight muscles at the expenseof egg production (Buckingham and Passoa 1985). It ispossible that, at relatively high densities or when foodquality is low, the female weevils switch to a disper-sive mode (Center and Durden 1986). Losses tonatural enemies may be less important, as few parasi-toids attack the weevils outside their native range(T. Center, pers. comm.) and, although birds havebeen seen to eat adults, adult weevils are generally notavailable to predators, hiding in the base of the peti-oles. However, the relatively quick success of biolog-ical control agents in western Mexico (T.D. Center,pers. comm.), is thought to be due to eliminating amicrosporidian infection which reduces the efficacy ofthe agents.
The most likely candidates for limiting the weevilsin stable conditions appear to be some form of larvalcompetition linked to plant nutritional status; pre-pupal mortality in silted areas; adult migration againlinked to low plant quality or parasitic burden i.e.microsporidians.
Other Modelling Approaches
The models described are designed to be general, andas such have simplifying assumptions; for example,that water hyacinth can be modelled as a population ofbiomass. However, to answer specific questions themodels may need more detail. Plant physiologicalmodels, based on the metabolic pool concept(Gutiérrez 1996), are being developed as part of theIMPECCA mycoherbicide project. These modelsinvestigate how different application strategies affectwater hyacinth population dynamics under differentenvironmental conditions. These questions necessitatea more detailed modelling approach, in particular onethat includes leaf dynamics. However, bothapproaches address the causes of variability in waterhyacinth infestations. If the models concur, the con-clusions should be independent of the modelling tech-nique used. The latest version of the model is availablefree at <http://www.agrsci.dk/plb/nho/hyacinth.htm>.
Discussion and Conclusions
In this section we draw the conclusions from the workon modelling water hyacinth, discuss how the additionof the weevils to the models affects the outcomes andwhy the models may not reflect the real situation.
Monitoring of water hyacinth should be carried outusing biomass and surface area covered, as these givethe best measure of the problem. Leaf length andnumber of leaves can be used as a surrogate, but theexact relationship may be very site specific. A simplelogistic model gives a good description of water hya-cinth growth. The parameterisation of the logisticmodel has highlighted several conclusions from otherstudies. Water hyacinth cannot survive at salinitiesabove about 0.2%. Water hyacinth can grow in watertemperatures of between 13 and 40°C and grows opti-mally at 30°C. High temperatures increase water hya-cinth mortality, but mortality does not increase at lowtemperatures without frost. Frost kills leaves and afterseveral days or a hard frost the meristem can bedamaged and the plant killed. Under constant condi-tions, water hyacinth shows a hyperbolic relationshipbetween water nutrient concentration and growth rate,with half-saturation co-efficients of between 0.05 and1 mg/mL for total nitrogen and between 0.02 and 0.1mg/mL for phosphates. However, complications meanthat the plant nutrient content is a more accurate guidewith a linear relationship between the percentagenitrogen in the leaves and the growth rate (Aoyamaand Nishizaki 1993). Water hyacinth growth rate isalso reduced by wave action, and in such environmentsit may persist only in sheltered regions or as part of amat. Natural enemies also limit water hyacinthgrowth. To test the model predictions, informationneeds to be collated on where water hyacinth has andhas not caused problems and how infestations develop.This would require historical data-sets possiblyincluding remote sensing.
When the models are adapted to include weevils, theprediction is for water hyacinth to be effectively erad-icated in all stable conditions. This qualitative predic-tion does not appear to be affected by refinements tomake the model more biologically realistic e.g. intro-ducing time delays corresponding to developmentaldelays or the introduction of stage structure. However,these predictions are altered in a dynamic situation e.g.frost and flooding. In order to test and refine themodel, information again needs to be collated onwhere control has and importantly where it has notbeen successful. One important aspect not included inthe models is spatial heterogeniety of attack. The adult
103
weevils are able to swim and are mobile, but in generalthey are sedentary and some plants may temporallyescape attack. It would be expected, therefore, that inlarger water bodies the reduction would be relativelyless than on a similar smaller water body. The model-ling work has thrown up a number of potentiallyimportant questions, the answers to which could bevery important in directing future control measures.
Under what conditions does water hyacinth remainat low levels? What limits the size of the weevil pop-ulations? Under what conditions are they limitedbelow a level that causes significant reductions in thewater hyacinth population? Do shallow or muddybanks provide refuge for water hyacinth plants by pre-venting N. eichhorniae from pupating? Does adultweevil migration prevent damaged water hyacinthfrom being eradicated? How does the interactionbetween water nutrient level, plant nutrient level andweevil damage affect the level of water hyacinth andweevils seen? How do the larvae compete when athigh densities? How are the weevils dispersed, and canthis account for failure in control? Why doesn’t N.eichhorniae work everywhere? These questions willbe addressed by refining the models, experimentationand more detailed investigation of field data.
Acknowledgments
We would like to thank Mic Julien, Ted Center, MartinHill, Peter Neuenschwander and Obinna Ajuonu forinvaluable discussions and allowing us access tounpublished data. Thanks also to Andy Wilby for con-structive comments.
References
Akbay, K.S., Howell, F.G. and Wooten, J.W. 1991. Acomputer simulation model of waterhyacinth and weevilinteractions. Journal of Aquatic Plant Management, 29,15–20.
Aoyama, I. and Nishizaki, H. 1993. Uptake of nitrogen andphosphate, and water purification by water hyacinthEichhornia crassipes (Mart.) Solms. Water Science andTechnology, 28, 47–53.
Bock, J.H. 1966. An ecological study of Eichhorniacrassipes with special emphasis on its reproductivebiology. In: Botany. Berkeley, University of California,Berkeley, 175.
Buckingham, G. and Passoa, S. 1985. Flight muscle and eggdevelopment in waterhyacinth weevils. In: Delfosse,E.S., ed., Proceedings of the VI International Symposiumon Biological Control of Weeds, Vancouver, Canada1984, 497–510.
Carignan, R. and Neiff, J.J. 1994. Limitation of waterhyacinth by nitrogen in subtropical lakes of the Paranáfloodplain (Argentina). Limnology and Oceanography,39, 439–443.
Caughley, G. and Lawton, J.H. 1981. Plant–herbivoresystems. In: May, R.M., ed., Theoretical ecology—principles and applications. Oxford, Blackwell, 132–166.
Center, T.D. 1987. Do waterhyacinth and otogeny affectintra-plant dispersion of Neochetina eichhorniae(Coleoptera: Curculionidae) eggs and larvae?Environmental Entomology, 16, 699–707.
Center, T.D. and Durden, W.C. 1986 Variation inwaterhyacinth/weevil interactions resulting fromtemporal differences in weed control efforts. Journal ofAquatic Plant Management, 24, 28–38.
Center, T.D. and Spencer, N.R. 1981 The phenology andgrowth of water hyacinth (Eichhornia crassipes (Mart.)Solms) in a eutrophic north-central Florida lake. AquaticBotany, 10, 1–32.
Debusk, T.A., Hanisak, M.D., Williams, L.D. and Ryther, J.H. 1981. Effects of seasonality and plant density on theproductivity of some freshwater macrophytes. AquaticBotany, 10, 133–142.
DeLoach, C.J. and Cordo, H.A. 1976. Life cycle and biologyof Neochetina bruchi, a weevil attacking waterhyacinth inArgentina, with notes on N. eichhorniae. Annals of theEntomological Society of America, 69, 643–652.
Ewel, K.C., Braat, L. and Stevens, M.L. 1975. Use of modelsfor evaluating aquatic weed control strategies. HyacinthControl Journal, 13, 34–39.
Forno, I.W. 1981. Effects of Neochetina eichhorniae on thegrowth of waterhyacinth. Journal of Aquatic PlantManagement, 19, 27–31.
Gopal, B. 1987. Water hyacinth. Aquatic plant studies 1.Amsterdam, Elsevier.
Guillarmod, A.J. and Allanson, B.R. 1978. Eradication ofwater hyacinth. South African Journal of Science, 74,122.
Gurney, W.S.C. and Nisbet, R.M. 1985 Fluctuationperiodicity, generation separation, and the expression oflarval competition. Theoretical Population Biology, 28,150–180.
Gurney, W.S.C., Nisbet, R.M. and Lawton, J.H. 1983. Thesystematic formulation of tractable single-speciespopulation models incorporating age structure. Journal ofAnimal Ecology, 52, 479–495.
Gutierrez, A. P. 1996 Applied population ecology: a supply–demand approach. New York, John Wiley and Sons,300p.
Gutiérrez, E., Franco, E.R., Gómez, E.U., Delgadillo, R.H.,and Jiménez, M.M. 2001. Biomass and productivity ofwater hyacinth (Eichhornia crassipes) and theirapplication in control programs. These proceedings.
104
Haller, W.T., Sutton, D.L. and Barlow, W.C. 1974. Effectsof salinity on growth of several aquatic macrophytes.Ecology, 55, 891–894.
Heard, T.A. and Winterton, S.L. 2000. Interactions betweennutrient status and weevil herbivory in the biologicalcontrol of water hyacinth. Journal of Applied Ecology,37, 117–127.
Imaoka, T. and Teranishi, S. 1988. Rates of nutrient uptakeand growth of the water hyacinth (Eichhornia crassipes(Mart.) Solms). Water Research, 22, 943–951.
Jayanth, K.P. and Visalakshy, P.N.G. 1990. Studies ondrought tolerance in the water hyacinth weevilsNeochetina eichhorniae and N. bruchi (Coleoptera:Curculionidae). Journal of Biological Control, 4, 116–119.
Julien, M.H., Griffiths, M.W. and Wright, A.D. 1999.Biological control of water hyacinth. The weevilsNeochetina bruchi and N. eichhorniae: biologies, hostranges, and rearing, releasing and monitoring techniquesfor biological control of Eichhornia crassipes. Canberra,Australian Centre for International AgriculturalResearch, ACIAR Monograph No. 60, 87p.
Julien, M.H., Harley, K.L S., Wright, A.D., Cilliers, C.J.,Hill, M.P., Center, T.D., Cordo, H.A. and Cofrancesco,A. F. 1996. International co-operation and linkages in themanagement of water hyacinth with emphasis onbiological control. In: Moran, V.C.H. and Hoffman, J.H.,ed., Proceedings of the IX International Symposium onBiological Control of Weeds, Stellenbosch, South Africa,273–282.
Knipling, E.B., West, S.H. and Haller, W.T. 1970. Growthcharacteristics, yield potential and nutritive content ofwater hyacinth. Soil and Crop Science Society of Florida,30, 51–63.
Lonsdale, W.M., Farrell, G. and Wilson, C.G. 1995.Biological control of a tropical weed: a population modeland experiment. Journal of Applied Ecology, 32, 391–399.
Lorber, M.N., Mishoe, J.W. and Reddy, P.R. 1984.Modelling and analysis of waterhyacinth biomass.Ecological Modelling, 24, 61–77.
Madsen, J.D. 1993. Growth and biomass allocation patternsduring waterhyacinth mat development. Journal ofAquatic Plant Management, 31, 134–137.
May, R.M. 1981. Theoretical ecology—principles andapplications. Oxford, Blackwell.
Mitsch, W.J. 1976. Ecosystem modelling of waterhyacinthmanagement in Lake Alice, Florida. EcologicalModelling, 2, 69–89.
Musil, C.F. and Breen, C.M. 1985a. The development fromkinetic coefficients of a predictive model for the growth
of Eichhornia crassipes in the field. I. Generating kineticcoefficients for the model in greenhouse culture.Bothalia, 15, 689–703.
— 1985b. The development from kinetic coefficients of apredictive model for the growth of Eichhornia crassipesin the field. IV. Application of the model to the VernonHooper Dam—a eutrophied South Africanimpoundment. Bothalia, 15, 733–748.
Newman, S. and Haller, W.T. 1988. Mineral deficiencysymptoms of waterhyacinth. Journal of Aquatic PlantManagement, 26, 55–58.
Nwankwo, D.I. and Akinsoji, A. 1988. Tolerance to salinityand survivorship of Eichhornia crassipes (Mart.) Solms.growing in a creek around Lagos. In: Oke, O.L.,Imevbore, A.M.A. and Farri, T.A., ed., Internationalworkshop on water hyacinth—menace and resource.Lagos, Nigeria, Nigerian Federal Ministry of Science andTechnology, 85–87.
Penfound, W.T. and Earle, T.T. 1948. The biology of thewater hyacinth. Ecological Monographs, 18, 448–472.
Reddy, K.R., Agami, M., D’Angelo, E.M. and Tucker, J.C.1991. Influence of potassium supply on growth andnutrient storage by water hyacinth. BioresourceTechnology, 37, 79–84.
Reddy, K.R., Agami, M. and Tucker, J.C. 1989. Influence ofnitrogen supply rates on growth and nutrient storage byWater Hyacinth (Eichhornia crassipes) plants. AquaticBotany, 36, 33–43.
Reddy, K.R., Agami, M. and Tucker, J.C. 1990. Influence ofphosphorus supply on growth and nutrient storage bywater hyacinth. Aquatic Botany, 37, 355–365.
Rees, M. and Paynter, Q. 1997. Biological control of Scotchbroom: modelling the determinants of abundance and thepotential impact of introduced insect herbivores. Journalof Applied Ecology, 34, 1203–1221.
Room, P.M. 1990. Ecology of a simple plant–herbivoresystem: biological control of Salvinia. Trends in Ecologyand Evolution, 5, 74–79.
Sahai, R. and Sinha, A.B. 1970. Contribution to the ecologyof Indian aquatics—I. Seasonal changes in biomass ofwater hyacinth (Eichhornia crassipes (Mart.) Solms).Hydrobiologica, 35, 376–383.
Visalakshy, P.N.G. and Jayanth, K.P. 1996. Effect of siltcoverage of water hyacinth roots on pupation ofNeochetina eichhorniae Warner and N. bruchi Hustache(Coleoptera: Curculionidae). Biocontrol Science andTechnology, 6, 11–13.
Watson, M.A. and Cook, G.S. 1987. Demographic anddevelopmental differences among clones of waterhyacinth. Journal of Ecology, 75, 439–457.
105
Current Strategies for the Management of Water Hyacinth on the Manyame
River System in Zimbabwe
G.P. Chikwenhere*
Abstract
The Manyame River System consists of the Manyame River and its tributaries, the Mukuvisi and Nyatsimerivers, which discharge water into man-made Chivero and Manyame lakes situated at 29 km and 33 km,respectively, from Harare. Water hyacinth is the predominant floating aquatic weed in the system. About 15years ago, the water was covered with various aquatic weeds namely: water hyacinth (35%), Pistia stratiotes(36.6%) and Myriophyllum aquaticum (1.7%). Ten years later, the estimated weed coverage was waterhyacinth (3%), P. stratiotes (0.3%), M. aquaticum (4%). Also, Azolla filiculoides had appeared in Lake Chivero,forming a 1% coverage. In 2000, the weed coverage on the system was 2.4, 0.8, 6,8 and 3.5%, respectively, and9.5% of the water surface of Lake Chivero had been invaded by Hydrocotyle ranunculoides. The control ofwater hyacinth was accomplished through chemical, biological and mechanical means, while the managementof P. stratiotes was accomplished mainly through classical biological control. The appearance of M. aquaticumand H. ranunculoides may have been facilitated by the absence of natural enemies, and reduced competitionfor space and nutrients by the previously dominant water hyacinth and P. stratiotes. In 2000, the ZimbabweAquatic Weed Management Committee was formed to manage the aquatic weed problem in a holistic manner.
ZIMBABWE has been involved in the biological controlof pests, including floating aquatic weeds, since the1950s (Chikwenhere 1994, 2001). The country hasbeen working through international collaboration andcooperation as a means to achieving sustainable clas-sical biological control technologies.
Following the First IOBC Global Working GroupMeeting for the Biological and Integrated Control ofWater Hyacinth, held in Harare, Zimbabwe, in 1998,Zimbabwe adopted a new coordinated approach in themanagement of the water hyacinth problems, with par-ticular attention to the Manyame River System thusincluding Chivero and Manyame lakes.
Formation of Zimbabwe Aquatic Weed Management Committee
The Zimbabwe Aquatic Weed Management Com-mittee was formed in July 2000 and it comprisesvarious stakeholders including government, universityand private sector representatives (Table 1). The com-mittee’s mandate is to identify water impoundmentsfor immediate chemical and biological control strate-gies for aquatic weeds, including chemical control ofnewly emerging water hyacinth seedlings, and to trainnational parks personnel in the proper use of sprayequipment. The committee is also charged with pre-paring an implementation plan for waterweed manage-ment in the short and medium terms. This would:* Plant Protection Research Institute, P.O. Box CY 550,
Causeway, Harare, Zimbabwe. Email: [email protected]
106
1. Identify areas for– immediate herbicide application in economically
important areas– areas requiring herbicide application if resources
permit– areas reserved for biological control.
2. Immediately spray germinating seedlings along thelake shore areas.
3. Train national parks staff in identification and massrearing of biological control agents for restocking inareas designated for biological control activities.
4. Train national park staff in calibrating sprayequipment, spray mixing and handling ofherbicides.The Zimbabwe Government provided the com-
mittee with an annual working budget of ZWD6m(US$125,000) for the year 2000 for aquatic weedcontrol along the Manyame River System. This systemhas a surface area of approximately 60 km2 and theestimated weed infestation level was below 5%.
Observations on the Pattern of Weed Infestation in Lake Chivero
(September 2000)
1. The major part of the weed biomass in the lake wasnot water hyacinth but the spaghetti weed,Hydrocotyle ranunculoides.
2. Water hyacinth mats were visible as isolatedpatches, brownish and severely weevil damaged,
surrounded by vigorously growing spaghettiweed.
3. On the Upper Manyame River System, spaghettiweed and parrot’s feather, Myriophyllumaquaticum, appeared to be the dominant species,outcompeting both water hyacinth and waterlettuce, Pistia stratiotes, for space, light andpossibly nutrients.
4. The most significant pocket of water hyacinth wasobserved at Tiger Bay, but even that was notexclusively water hyacinth, as spaghetti weedconstituted a larger proportion of the weedbiomass.
5. Other significant pockets of water hyacinth werejust below the Lake Chivero Spillway but thesewere less then 10 cm tall and even in these areas,spaghetti weeds constituted a significantproportion of the biomass.
6. Along the northern shore of Lake Chivero,spaghetti weed formed a continuous fringeextending about 3–4 m from the shoreline into thewater. After the spaghetti weed fringe, the waterhyacinth formed another belt of about 1 m. Onaverage, the entire weed belt around the lake wasabout 4 m, approximated 6.5% of the surface area.The current weed cover remains far less than the35% previously recorded on the lake in the 1980sbefore the release of the Neochetina weevils(Table 2).
Table 1. Composition of Zimbabwe Aquatic Management Committee
Institution/organisation Representative Area of interest
Harare City Council Mr T. Mafuko Urban water supply and effluent disposal systems
Ministry of Agriculture, Plant Protection Research Institute
Dr G. P. Chikwenhere Biological control using insects
Agruricura: Agricultural Chemical Company Mr A. Brent Chemical control
University of Zimbabwe, Department of Biological Sciences
Prof. B. Marshal Use of lake waters and water hyacinth control
Ministry of Environment and Tourism, Department of National Parks and Wildlife
Mrs S. Mutsekwa Fish ecology and committee chairperson
Commercial Farmers Union Unadale Farm Tobacco and mixed cropping systems
Lazy River Fisheries Mr G. Manuwere Fishing
Zvevanhu Fisheries Mr B. Chidawanyika Fishing
107
Table 2. Comparisons on previous and present weed cover on the Manyame River System in Zimbabwe
Locality/area Weed species Estimated surface area covered (%)
Year observed
Common name Scientific name
Upper Manyame Water hyacinth Eichhornia crassipes 60.0 1986
Water lettuce Pistia stratiotes 35.0
Parrot’s feather Myriophyllum aquaticum 0.1
Red water fern Azolla filiculoides –
Lake Chivero Water hyacinth Eichhornia crassipes 35.0
Water lettuce Pistia stratiotes 40.0
Spaghetti weed Hydrocotyle ranunculoides –
Red water fern Azolla filiculoides –
Mukuvisi River Water hyacinth Eichhornia crassipes 10.0
Water lettuce Pistia stratiotes 35.0
Parrot’s feather Myriophyllum aquaticum 5.0
Red water fern Azolla filiculoides –
Upper Manyame Water hyacinth Eichhornia crassipes 4.0 1996
Water lettuce Pistia stratiotes 0.5
Parrot’s feather Myriophyllum aquaticum 4.5
Red water fern Azolla filiculoides 0.1
Lake Chivero Water hyacinth Eichhornia crassipes 5.0
Water lettuce Pistia stratiotes 0.5
Spaghetti weed Hydrocotyle ranunculoides 3.5
Red water fern Azolla filiculoides 1.0
Mukuvisi river Water hyacinth Eichhornia crassipes 0.8
Water lettuce Pistia stratiotes 0.1
Parrot’s feather Myriophyllum aquaticum 4.5
Red water fern Azolla filiculoides 2.7
Upper Manyame Water hyacinth Eichhornia crassipes 3.5 2000
Water lettuce Pistia stratiotes 1.5
Parrot’s feather Myriophyllum aquaticum 4.5
Red water fern Azolla filiculoides 5.5
Lake Chivero Water hyacinth Eichhornia crassipes 3.5
Water lettuce Pistia stratiotes 1.0
Spaghetti weed Hydrocotyle ranunculoides 9.5
Red water fern Azolla filiculoides 0.5
Continued on next page
108
Discussion
There is still an overall reduction in water hyacinthbiomass on Lake Chivero, as has been observed byChikwenhere and Phiri (1999). The rapid increase ofspaghetti weed and other floating macrophytes, partic-ularly in Lake Chivero, may have been facilitated bythe absence of natural control enemies, as well as thecompetitive ability of the weed in the presence ofwater hyacinth plants heavily stressed by Neochetinaweevils and water lettuce stressed by Neohydronomusaffinis weevils. Furthermore, nutrients previously uti-lised by large mats of water hyacinth and water lettucebecame available and may have contributed to therapid proliferation of the other weeds. At the sametime, accelerated succession of spaghetti weedseemed to be favoured by dead biomass of water hya-cinth plants, which provided a suitable substrate forthe new invader.
The committee has recommended chemical controltargeted only on spaghetti weed, but it remainedunclear whether water hyacinth will show an upturnonce the spaghetti weed levels have gone down.
Herbicidal and mechanical control have tradition-ally been the methods of choice, but the rising cost ofherbicides, and environmental concerns, prompted aninvestigation of the possibility of using biological con-trol. This aspect of research was initiated during the1980s (Chikwenhere 1994), and during the 1990s sig-nificant improvement in water hyacinth and waterlettuce control was achieved, especially in theManyame River System.
Observations during the past decade have shownthat, while effective control of water hyacinth has been
achieved, spagetti weed, A. filiculoides and M. aquat-icum have increased, and replaced water hyacinth andwater lettuce in many places. Studies of the competi-tive interactions of the weeds may assist developmentof holistic weed management strategies.
In Zimbabwe, lakes and reservoirs provide waterfor domestic use and for irrigation of commercialagriculture and small-scale farming communities.They have aesthetic value and are an important sourceof foreign exchange earning through tourism. TheAquatic Weed Control Management Committee wasformed in 2000 to oversee management of aquaticweeds. The committee will draw up an implementa-tion plan to accomplish aquatic weed control in aholistic and prioritised approach. It will also beresponsible for monitoring, evaluation and impactassessment of the control measures.
References
Chikwenhere, G.P. 1994. Biological control of waterhyacinth (Eichhornia crassipes)—results of a pilot study.FAO Plant Protection Bulletin, 42, 185–190.
— 2001. Biological control in the management of waterhyacinth and other macrophytes in Zimbabwe—a review.Workshop on the conservation and management ofwetlands in Zimbabwe, 12–14 February 1997, Harare,Zimbabwe.
Chikwenhere, G. P. and Phiri, G. 1999. History and controleffort of water hyacinth, Eichhornia crassipes on LakeChivero in Zimbabwe, In: Hill, M.P., Center, T.D. andJulien, M.H., ed. Proceedings of the First IOBC GlobalWorking Group Meeting for the Biological andIntegrated Control of Water Hyacinth, Harare,Zimbabwe, November 1998, 91–100.
Mukuvisi River Water hyacinth Eichhornia crassipes 0.2
Water lettuce Pistia stratiotes 0.1
Parrot’s feather Myriophyllum aquaticum 6.5
Red water fern Azolla filiculoides 2.4
Table 2. (Cont’d) Comparisons on previous and present weed cover on the Manyame River System in Zimbabwe
Locality/area Weed species Estimated surface area covered (%)
Year observed
Common name Scientific name
109
Biomass and Productivity of Water Hyacinth and Their Application in Control Programs
E.L. Gutiérrez, E.F. Ruiz, E.G. Uribe and J.M. Martínez*
Abstract
Water hyacinth is controllable if management programs take account of plant dynamics and factors that influenceplant behaviour. The project for reclamation of water bodies in Mexico considered water hyacinth standing crop,coverage and growth. The method proposed served to characterise the initial population and monitor the controlprocess. The growth model used was reliable in predicting the effective reduction in the weed in response tocontrol pressure. Change in growth over an annual cycle was characterised by a sigmoid curve. The maximumrelative percentage growth rate was 9.34%, with a duplication time of 7.4 days from April to June. During winter,growth decreased by up to 90%. In a dam, 144 t/ha/year of dry matter was produced, characteristic of water plantswith a high nutrient content. The water hyacinth population can be reduced by 90% through water levelmanagement and mechanical destruction. For example, approximately 3600 t/day was removed over 181 days toreduce the infestation to manageable levels. Physical, chemical and biological methods are used to maintain theselevels, but input of urban and industrial contaminants must be controlled for long term rehabilitation.
OUTBREAKS of aquatic plants is the result of changesin the physical, chemical and biological conditionsbrought about by the uncontrolled flow of nutrientsfrom urban, agricultural and industrial centres and insilt eroded from watersheds (Gutiérrez et al. 1994).
Water hyacinth is successful owing to its life cycleand survival strategies that have given it a competitiveedge over other species. Its adaptability to little com-peted ecological conditions make eradication of thisplant virtually impossible and control extremely diffi-cult (Gutiérrez et al. 1996). In Mexico, more than40,000 ha were infested and specific management pro-grams were needed. The Aquatic Weed ControlProgram (AWCP) was created in 1993 to combat theexcessive presence of the weed in the nation’s water-courses.
The aims of the AWCP included to:• reduce the weed to a manageable level and maintain
this level through a maintenance programdeveloped for the body of water;
• use methods most suitable to ecosystem and wateruses;
• formulate an integral watershed program which willinclude the control and maintenance operations; and
• establish biological control using insects and fungi.Under a national program to control the water hya-
cinth, guidelines to deal with the related environ-mental, social, technical and economic factors, andspecific strategies to reduce coverage were developed.The environmental factors included the identificationof the characteristics of the affected areas and the con-sequences of the proposed treatments. The socialaspects embraced the stimulation of user awareness ofthe importance of water quality, the creation of organ-isations to coordinate user-sponsored control activi-ties, and the awakening of the community identity.
* Instituto Mexicano de Tecnología del Agua, PaseoCuauhnáhuac 8532 colonia Progreso, Jiutepec, Morelos62550, Mexico. Email: [email protected]
110
Basic to all are the technical and economic aspectswhich make the activities feasible and operational(Gutiérrez et al. 1996).
Most of the water hyacinth control methods havebeen used in Mexico; harvesting by hand and machine,mechanical crushing, and treatment with herbicidesand biological agents. Experience indicated that waterhyacinth is controllable if the process takes account ofplant dynamics and factors that influence plant behav-iour. The project considered biomass, cover andgrowth to be important. The objective of this work wasto characterise the initial population of the weed toassist with and assess control programs.
Materials and Methods
The best control strategy is that which reduces thebiomass of the water hyacinth in a reasonable time andat an acceptable cost, i.e. the use of one or severalcontrol methods that effectively reduce the amount ofplants faster than its natural reproduction, without neg-atively affecting the ecosystem. Even though there aremany interacting variables and components, the behav-iour of the plants is one of the most important factors.In terms of weed control, plant behaviour can bestudied through three parameters: biomass, infestationlevel (surface area covered), and growth rates. Thesefactors vary in time and space and are site-specific.
Biomass
The biomass is defined as the amount of weed massin a particular area or volume. The effect of a waterhyacinth population over the water ecosystem dependson this characteristic. The excessive increase inbiomass is an indication of an increment in the energyconversion rate caused by the availability of resources.Water hyacinth is an example of an exotic plant thatcompetes effectively for the space.
Concerning mechanical control, Hutto and Sabol(1986) mention that the effectiveness of a croppingsystem depends mainly on the standing crop becausethat determines a machine’s movement rate throughouta work site and the number of loads that can be trans-ported. Biomass of plants also influences the efficacyof herbicides. In practice, it seems unlikely that 100%control of water hyacinth using herbicide can beobtained when it grows in heavy infestations becauseadjacent plants screen one another (Gutiérrez 1993).These considerations show the need to measure the
standing crop of water hyacinth in the infested waterbodies where a control program is to be established.
Plant biomass was obtained by weighing samplesfrom the field and estimating the weight of the popu-lation. One square metre samples were collected,drained for 5–7 minutes, and weighed using a 50 kg(± 1 kg) scale. Sub samples of 1 kg were dried to con-stant weight and weighed.
The number of samples per sampling, N, was deter-mined according to Madsen (1993):
where s is the standard deviation and x is the mean.
Cover
Cover was defined as the space covered by the weedas seen from above (Brower and Zar 1977). To esti-mate cover on small water bodies, estimates weremade by mapping the infestation, at different times,while standing on a predetermined, elevated set-point.The area covered was then determined for each date bycomparing the mapped infestation with the knownarea of the water body. The area was used with the esti-mated weight per unit area to calculate total biomass.
Landsat-TM satellite images were used to estimatecover on large water bodies. In the images a ‘falsecolour’ compound is generated through the SatelliteImage Automatic Detection System (SIADIS), byhighlighting areas and combining bands 1, 2, 3, 4, 5and 7, using blue, green and red filters, respectively.
Growth
Weed growth was determined from the weightincrease of the water hyacinth mass per area unit andper time unit, i.e. its productivity (Westlake 1963).
Quantification of rate of growth is important forcontrol. The rate is affected by factors such as nutri-ents, climate, space and compaction.
To measure growth four 1 m2 compartments wereinstalled into the edges of water hyacinth mats. Allmaterial was removed from inside the compartments.One kg of selected ramets (healthy, undamaged, with3–5 leaves, of uniform size and weighing 30 to 45 geach) was placed into each compartment and allowedto grow. After 30, 60 and 90 days, the wet weight ofthe 2 m2 was obtained using the same procedure forbiomass determination described above.
Ns
x=
×( )2
20 1.
(1)
111
For comparison purposes between sites and otherdata obtained in different water bodies, the daily rela-tive growth rate (RGR%) and the biomass doubling-time (DT) were calculated according to Mitchell(1974, cited in Sastroutomo et al. 1978):
where Xo is initial weight and Xt is weight after t days.The three parameters (biomass, cover and growth
rates) were obtained in seven water bodies whose maincharacteristics are shown in Table 1. These reservoirswere classified mesotropic if phosphorus concentra-tions were between 10 and 35 mg/m3 or eutrophic ifphosphorus concentrations were 35 to 100 mg/m3
(Vollenweider 1983).
Results and Discussion
The highest biomass average was 49.6 (2.79) kg/m2,and a maximum value of 76 (4.27) kg/m2, occurred inCruz Pintada Dam. This was the smallest dam studiedand had the highest level of compaction. In general,these values are similar to those obtained in other partsof the world, except for a value of 5.96 kg/m2 dryweight observed in Jaipur, India (Trivedy 1980).
Maximum cover generally occurred when thesurface area was smallest, and consequently, storagevolume lowest. The extraction of water from the damsstranded a great part of the water hyacinth on bankswhere some died of desiccation. Other plants recov-ered when water levels increased.
The purpose of measuring water hyacinth growth wasto know the relative behaviour of the biomass in anenvironment that is generally favourable for itsincrease. The form of this increase and its mathematicalrepresentation can be used as a starting point to plan acontrol program.
It will not be possible to reduce plant infestationwhile the removal rate of the biomass, either by har-vesting, crushing or another procedure, is less than itsgrowth recovery rate.
Table 3 shows the weight changes measured atRequena Dam. Data for location (a) in Table 3, themost comprehensive data set, are also presented asFigure 1.
This ratio showed a growth approximated to thelogistic equation 4.
where:Wt is wet weight for each determined time (kg/m2);r is growth rate per day;K is growth limit value of the population or loadcapacity (kg/m2);t is time; anda is an integration constant defining curve positionin relation to its origin.When supposing a growth of this type, the parameters
r and a can be calculated, and the logistic equationtransformed into its rectilinear form (equation 5):
The results of this exercise are shown in Table 3(a).Even though the correlation of the points was veryhigh (0.986), a significance test of the regression wascarried out according to Zar (1974). This test rejected,with a probability higher than 99%, the possibility thatthe points over the straight line are adjusted by chance.The same test was carried out to the data in Table 3(b)and 3(c) resulting also in the rejection of the possibilitythat the points are adjusted to a straight line by chance,except that for both cases the reliability level was 95%.
It is accepted that the water hyacinth growth is closeto logistic growth. Sato and Kondo (1983) establishedthat the biomass increase (fresh weight per surfaceunit) closely approximates the logistic equation; andDel Viso et al. (1968) demonstrated that the annualgrowth cycle of this plant in Argentina can be repre-sented by a sigmoid curve.
Reddy and Debusk (1984), in growth evaluationswith plants cultivated in a pond with unlimited nutri-tional conditions, determined the growth characteris-tics of water hyacinth in the central part of Florida,USA. They obtained a growth curve characterised bythree phases: 1. a delay phase followed by exponentialgrowth; 2. a linear growth phase, and 3. a slow expo-nential growth phase. These characteristics are verysimilar to the results obtained in this study, wherebehaviour was measured directly in the field.
We considered that the carrying capacity of thesystem (K), was reached during the periods whenmaximum biomass was obtained: 51 kg/m2 for July to
RGRX
Xt%
ln
ln= ( )
0
100 (2)
DTRGR
= ln2(3)
WK
et a rt=
+ −1(4)
WK W
Wa rtt = − = − (5)
112
February (Table 3a), 51 kg/m2 for December to March(Table 3b) and 55 kg/m2 for April to June (Table 3c).These values are not shown in the respective tablesbecause the values shown are averages. Reddy andDebusk (1984) suggested that the water hyacinthgrowth cycle was complete when the maximumdensity of plants was reached and therefore an addi-tional significant biomass increment was notobserved. They found a maximum biomass close to
2,300 g /m2 in dry weight, while in this study a rangeof 2,101–3,916 g /m2 was estimated.
The r and K parameters from the logistic equationprovide an objective comparison between differentwater systems. They also provide a foundation for aprospective model of the water hyacinth behaviour ona water body, as influenced by different rates ofbiomass removal.
Table 1. Characteristics of the seven water bodies under study (modified from Bravo et al. 1992)
Parameter Chairel Lagoon
Cruz. Pintada Dam
Sanalona Dam
Solís Dam Requena Dam
Endhó Dam Valle de Bravo Dam
North latitude 22° 16' 18° 26' 24°48' 20° 04' 19° 57' 20° 04' 19° 21'
West longitude
97° 54' 99° 01' 107° 09' 100° 35' 99° 18' 99° 20' 100° 11'
Altitude(m)
0 1,011 135 1,880 2,110 2,018 1,830
Climate Hot Hot Hot Temperate Temperate Temperate Temperate
Temperature (°C)
24.3 22.0 24.4 21.5 15.4 17.0 18.1
Precipitation (mm)
1,096 800–1,000 814 734 553 609.4 1236.9
Surface area(km2)
38.790 0.100 24.000 57.02 5.4 8.43 17.3
Volume (’000 m3 )
28,794 400 473,000 794,000 30,300 107,900 300,000
Depth(m)
1 a 3 4 19.7 14.0 5.0 15 19.4
Trophic level mesothropic eutrophic mesothropic eutrophic eutrophic eutrophic mesothropic
Table 2. Weight per m2, surface area covered and total biomass of seven water bodies in Mexico (modified fromBravo et al. 1992)
Reservoir Standing crop wet (dry) weight
Cover Total biomass(t)
Average(kg/m2)
Maximum(kg/m2)
Average(ha)
%
Chairel Lagoon 39.5 (2.22) 50.5 (2.84) 376 10 148,520
Cruz Pintada Dam 49.6 (2.79) 76 (4.27) 7.5 75 3,720
Sanalona Dam 42.6 (2.39) 57 (3.20) 790 33 336,540
Solís Dam 38.8 (2.18) 63 (3.54) 3,378 59 1,310,664
Requena Dam 35.74 (2.0) 51 (2.87) 498 70 175,803
Endhó Dam 33.5 (1.88) 51 (2.87) 818 80 220,000
Valle de Bravo Dam 45.7 (2.57) 67 (3.76) 109 6 50,00
113
Table 3. Comparitive studies at Requena Dam at three locations (a), (b) and (c).
Date Time (days) Biomass (kg/m2) Logistic equation parameters
Doubling time (days)
(a) 16-07-8614-08-8617-09-8613-10-8618-11-8610-12-8619-01-8717-02-87
0296389
125147187216
0.252.7015.426.039.045.050.050.5
a = 4.7073r = 0.0499K = 51 kgCorr. = 0.9860Reliability:greater than 99%
8.2–8.45
(b) 10-12-8619-01-8717-02-8717-03-87
0406997
0.2500.5630.6751.288
a = 5.2780r = 0.0162K = 51 kgCorr = 0.9838Reliab. = 95%
2.03–34.66
(c) 28-04-8712-05-8712-06-8730-07-87
0144893
1.03.7
22.053.5
a = 3.2746r = 0.0722K = 55 kgCorr = 0.9598Reliab. = 95%
9.34–7.42
0
10
20
30
40
50
60
Bio
mas
s (k
g/m
2 )
0 10 20 30 40 50 60 70 80 90 100
110
120
130
140
150
160
170
180
190
200
210
216
220
Biomass est.
Biomass obs.
dw/dt
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
dw
/dt
t (days)
Figure 1. Weight changes for water hyacinth at Requena Dam
114
Regarding the growth rate, in Florida an averagerate of 52 g/m2/day was observed during June andJuly, with a maximum value of 64 g/m2/day. At theRequena Dam, rates of 59.1 and 60.4 g/m2/day wereestimated for the July to February and April to Juneperiods. The growth rates in both studies were calcu-lated from the slope of the growth curve, adjusted bysquare minimums. If we consider an average growthrate of 0.551 tonne/ha/day, during the growth season(April to November, 244 days), approximately 134.4tonne/ha/year can be produced in the dam. Westlake(1963) qualified the water hyacinth as a very produc-tive plant. From data of Louisiana, USA and the Nile,Africa, he estimated that if this species grows undergood conditions, with a good density and withoutspace limitations and a continuous predominance ofyoung plants, it can produce as much as 110–150tonnes of organic matter/ha/year, a value very close tothe value estimated in this study.
A wide range of values for the productivity of thisplant has been registered in the literature. These valueshave been calculated in different ways (Gopal 1987).Knipling et al. (1979) estimated that the annual pro-duction can be as high as 269 t/ha. Boyd (1976, cited inGopal 1987) obtained an average productivity of 194kg/ha/day in an enriched pond. Wooten and Dodd(1976) and Yount and Crossman (1970) determined,respectively, a daily productivity of 290 and 540 kg/ha; the latter value corresponds to a eutrophic lake.Singh et al. (1984) and Wolverton and McDonald(1979, cited in Gopal, 1987), estimated a daily produc-tion of biomass of 26 and 72 g/m2, the latter for waste-water effluent.
This information shows that water hyacinth has awide productivity range. However, the values that arecloser to those obtained in this study are similar tothose generated in waters with high content of nutri-ents, wastewater effluents and eutrophic water bodies.
The determinations made are considered a goodapproximation to the net primary productivity of thisspecies. Corrections were not made for death, diseaseand herbivory. Westlake (1963) indicate that unlessherbivory is visually obvious, it is probably not impor-tant. Herbivory was not observed in this study.
Rates of loss due to natural plant death vary fromplace to place. Because leaf production is constant andproportional to leaf mortality, each mature shoot main-tains a relatively constant number of leaves (Center etal. 1984; Center 1987). Generally, the losses are nothigher than 2–10% of the maximum biomass (Harper
1918; Borutskii 1950; Westlake 1965; all cited bySculthorpe 1967). In tropical and subtropical habitats,mortality occurs throughout the year, usually as muchas new material is produced, so that the biomassremains more or less constant (Sculthorpe 1967).
For comparison purposes, it is appropriate to calcu-late the relative percentage growth rate (RGR%) andthe DT of the water hyacinth biomass. The RGR% andthe DT were calculated for the first measurement ofeach experimental lot. They are shown in Table 3.
The daily RGR% was between four to five timesgreater in summer and spring than in winter, resultingin a shorter DT of the biomass. These results aresimilar to those obtained by Sastroumoto et al. (1978)who determined in Chiba, Japan, that the RGR% andthe DT of the water hyacinth was five times higher andfour times faster in summer than in winter. Theseauthors observed that if fertiliser (10 kg N, P, and K)were added, the RGR became eight times higher andthe DT five times shorter. We concluded that the dif-ferences between spring–summer and winter found inthe Requena Dam are the result of differences in waterquality rather than in temperature.Table 4 showsRGR%, DT, K and r values estimated in the otherwater bodies that were assessed. In Mexico, thehighest RGR% value obtained was in Requena Dam(9.34%). Higher values were obtained in Florida, 12%(Cornwell et al. 1977) and in the Sudan, 11.8% (Pettet1964), both in summer and under natural conditions.Growth of water hyacinth is influenced by a number offactors. However, in Mexico its growth varies across arange from 1.07 to 12%.
We did not determine the cause of the variation ingrowth. The most important factors that influence waterhyacinth growth are known to be nutrient availabilityand temperature. However, those factors do not explainwhy in places lower than expected rates of growthoccurred, for example, Solís Dam or Sanalona Dam.
Control model
The logistic model expressed in equation 4 is theresult of the differential equation (equation 6) that,once integrated, represents the growth characteristicsfound in the Requena Dam:
dW
dtrW
r
kW= − 2 (6)
115
However, to consider the effect of biomassremoval, it is necessary to include the correspondingterm into equation 6. Thus the expression is (Romeroet al. 1989);
where R is the amount of water hyacinth that can beremoved (kg/day) and A is the reservoir area (m2)covered with water hyacinth.
This model presupposes that the biomass of the plants(W) is distributed evenly in the surface of the waterbody. The growth rate (r) is proportional to the densitywhen density is low; as density increases, the growthrate diminishes slowly until the maximum biomass (K)is reached. Normally the biomass in K (load capacity),that is, the asymptote in equation 4, stays withoutapparent changes, which can be caused when impactinga control process or removal of the weed, included in themodel with the term – [R/A].
This model consists of four components:
1. a variable growth rate (r) determined by the amountof initial biomass;
2. a measurement of the population size (W);
3. a measure of the limiting factor of growth (– [r/k]W2); and
4. a measurement of the biomass loss (– [R/A]).
Romero (1989) deduced from equation 7 that thispoint can represented as in equation 8.
with R* in kg/day.This expression is of practical usefulness because it
allows us to mathematically predict the total biomassbehaviour of water hyacinth in the Requena Dam at itsmaximum infestation and the effect exercised by thecrusher and other actions for its decline.
Thus, if we have:
Requena Dam area A= 4,928,300 m2
Growth rate r = 0.049 kg/kg/dayLoad capacity K = 51.0 kg/m2Removal capacity R
Substituting these data in equation 8 we obtain:
R* = ArK/4 = 3,080,000 kg/day = 3,080 t/day
If the actual rate of removal was lower than R* thecover would never be reduced. However, if low initialbiomass was present, reduction of cover (or biomass)would be possible. The model greatly depends on theinitial density of the water hyacinth, i.e. the biomassper m2 when population removal begins.
If the removal rate was greater than R*, reductionin cover would be achieved. Figure 2 shows thebiomass behaviour in each of the seven dams under aparticular control level. For example, Figure 2eshows the decline in water hyacinth in Requena Dam,if 3,600 tonne/day was removed. A theoretical zerobiomass value would be reached at about 200 days.
Table 4. The relative growth rate (RGR), doubling time (DT), carrying capacity (K) and intrinsic rate of increase (r)for water hyacinth in seven water bodies of the Mexican Republic (modified from Bravo et al. 1992).
Water body Relative growth rate (RGR)
(%)
Doubling time (DT)(days)
Load capacity K
(kg/m2)
Intrinsic growth rate r
(1/days)
4.45
Chairel 1.49 15.58 46.1 0.038
Cruz Pintada 1.07 46.53 60.7 0.152
Sanalona 2.66 64.56 49.0 0.0110
Solís 4.45 26.07 51.1 0.0274
Requena Summer 8.20 8.45 51 0.049
Winter 2.03 34.60 51 0.016
Spring 9.34 7.42 55 0.072
Endhó 7.07 9.9 55 0.065
Valle de Bravo 1.93 13.0 47 0.052
dW
dtrW
r
kW
R
A= − −2 (7)
RArK* =
4(8)
116
5
10
15
20
25
30
35
40
45
10
20
30
40
50
60
0
0
5
10
15
20
25
30
35
40
45
0
5
10
15
20
25
30
35
40
45
Bio
mas
s (k
g/m
2 )Bi
om
ass
(kg
/m2 )
Bio
mas
s (k
g/m
2 )Bi
om
ass
(kg
/m2 )
Time (days)
Time (days) Time (days)
Time (days)
R = 3000 t/day
R = 3000 t/day
R = 500 t/day
R = 17,000 t/day
0 10 20 30 40 50 60 70 80 90 100 1 2 3 4 5 6 7 8 9 1100
0 50 100 150 200 0 50 100 150 200 250
10
(a)
(c)
(b)
(d)
Figure 2. The estimated rate of removal and the duration to achieve a theoretical 100% removal of water hyacinth from each of seven dams; (a) Chairellagoon; (b) Cruz Pintada dam; (c) Sanalona dam; (d) Solís dam
117
0
5
10
15
20
25
30
35
40
0
5
10
15
20
25
30
35
40
0
510
1520
25
3035
4045
50
Bio
mas
s (k
g/m
2 )
Bio
mas
s (k
g/m
2 )
Bio
mas
s (k
g/m
2 )
R = 3600 t/day
R = 1000 t/day
R = 8000 t/day
Time (days)
0 50 100 150 200 250
Time (days)
0
Time (days)
0 20 40 60 80 100 120
50 100 150 200 250
(e) (f)
(g)
Figure 2. (cont’d) The estimated rate of removal and the duration to achieve a theoretical 100% removal of water hyacinth from each of seven dams; (e) Requena dam; (f) Endh dam; (g) Valle de Bravo dam
118
This removal logistic model was used to estimate therate of removal required for a number of water bodiesand was successfully applied.
The model assumes a uniform distribution of thewater hyacinth over the whole surface of the water andassumes a constant rate of removal. In spite of theselimitations the model allowed us to predict thebiomass changes and was a useful tool in planning forthe control of this weed.
The weed control program included 15 water bodies(Gutiérrez et al.1996). Aquatic weeds were removedby mechanically crushing in Requena Dam, EndhoDam and Valle de Bravo Dam (100% clear) and bymechanically harvesting in Chairel Dam and CruzPintada Dam (30% clear). Chemical control was usedin Solís Dam, where 100% clean-up was obtained.
It is impossible to remove all water hyacinth due togermination of seeds and regrowth and so manage-ment strategies to keep the weed at lower infestationlevels are required. Biological control is being usedand Neochetina adults are produced and released inseveral water bodies every month. Between April1994 and August 1998, 85,000 adults were released in15 water bodies including Requena Dam, Endho Damand Cruz Pintada Dam. Numerous feedings scars wereobserved on almost all plants and no substantial reduc-tion in plant size, wet weight or number of plants persquare metre was observed 4 years after initial releasesof Neochetina species. Plant reproduction may beoccurring much more rapidly than the weevils caninflict damage (J.M. Martinez, pers. comm. 2000).
Limitations of Neochetina in control of water hya-cinth were recognised by Perkins (1973), DeLoach andCordo (1976) and Perkins (1978). The effectiveness ofbiological control may be improved by the use of addi-tional agents (Charudattan 1986; Forno and Cofranc-esco 1993; Martinez et al. 2001). In some locations theNeochetina weevils are effective by themselves (Julien2001). Studies have begun in Mexico to determineindigenous species of pathogens and to evaluate howthe most promising of these may be applied as biolog-ical herbicides in areas where Neochetina is present, inorder to enhance the control effect.
References
Bravo, I.L., Sánchez, Ch.J., Olvera, V.V. and Gutiérrez, L.E.1992. Simulación de control mecánico de lirio acuático(Eichhornia crassipes). In: Sociedad Mexicana deIngeniería Sanitaria y Ambiental, ed., A.C. Memorias deVIII Congreso Nacional, Acciones para un Ambiente
Limpio, 22–25 septiembre 1992. Cocoyoc, Morelos,Mexico, Capitulo V.
Brower, E.J. and Zar, H.J. 1977. Field and laboratorymethods for general ecology. Iowa, USA, W.C. BrownCo. 174p.
Charudattan, R. 1986. Integrated control of waterhyacinth(Eichhornia crassipes) with a pathogen, insects, andherbicides. Weed Science, 34, 26–30.
Center, T.D., Durden, W.C. and Corman, D.A. 1984.Efficacy of Sameodes albiguttalis (Warren)(Lepidoptera: Pyralidae) as a biocontrol ofwaterhyacinth. U.S. Army Corps of Engineers AquaticPlant Control Program Technical Report A-84-2.
Center, T.D. 1987. Do water hyacinth leaf age and ontogenyaffect intra-plant dispersion of Neochetina eichhorniae(Coleoptera: Curculionidae) eggs and larvae?Environmental Entomology, 16, 699–707.
Cornwell, DA., Zoltek, J. Jr., Patrinely, C.D., Furman, S. de.T. and Kim, F.J. 1977. Nutrient removal by waterhya-cinths. Journal of the Water Pollution Control Federation,49, 57–65.
DeLoach, C.J. and Cordo, H.A. 1976. Ecological studies ofNeochetina bruchi and N. eichhorniae on waterhyacinthin Argentina, Journal of Aquatic Plant Management, 14,53–59.
Del Viso, R.P., Tur, N.M. and Mantovani, V. 1968.Estimación de la biomasa de hidrófitos en CuencasIsleñias del Paraná Medio. Physis, Buenos Aires, 28,219–-226.
Forno, I.W. and Cofrancesco, A.F. 1993. New frontiers inbiocontrol. Journal of Aquatic Plant Management, 31,222–224.
Gopal, B. 1987. Water hyacinth. Aquatic Plant, Studies 1.Amsterdam, The Netherlands, Elsevier Science, 471p.
Gutiérrez, L.E. 1993. Effect of glyphosate on differentdensities of water hyacinth. Journal of Aquatic PlantManagement, 31, 255–257.
Gutiérrez, L.E., Saldaña F.P. and Huerto, R.D. 1994.Informe final del proyecto: Programa de MalezasAcuáticas (Procina). Instituto Mexicano de Tecnologiadel Agua (Informe interno), México, 114p.
Gutiérrez, L.E., Huerto, R.D., Saldaña, F.P. and Arreguin, F.1996. Strategies for waterhyacinth (Eichhorniacrassipes) control in Mexico. Hidrobiologia, 340, 118–185.
Hutto, D.T. and, Sabol, M.B. 1986. Application of harvest,mechanical, bioaccumulation models, as an operationalaquatic macrophyte management decision tool. In: NorthAmerican Lake Management Society, Fifty AnnualConference and International Symposium on AppliedLake and Watershed Management, 13–15 November1985. Lake Geneva, Wisconsin, 2, 267–279.
Julien, M.H. 2001. Biological control of water hyacinth witharthropods: a review to 2000. In: Julien, M.H., Hill, M.P.,
119
Center, T.D. and Jianqing, D., ed., Proceedings of the 2ndIOBC Water Hyacinth Working Group Meeting, October2000, Beijing.
Knipling, E.B., West, S.H. and HaIler, T.W. 1979. Growthcharacteristics, yield potential, and nutritive content ofwaterhyacinths. Soil and Crop Science Society of Florida,30, 198–212.
Madsen, J.D. 1993. Biomass techniques for monitoring andassessing control of aquatic vegetation. In: Madsen, J.D.and Bloomfield, A.J., ed., Aquatic vegetationquantification for lake management. Lake and ReservoirManagement, 7, 141–154.
Perkins, B.D. 1973. Potential for waterhyacinthmanagement with biological agents. Proceedings of theTall Timbers Conference on Ecology, Animal Controland Habitat Management, 4, 53–64.
Perkins, B.D. 1978. Enhancement of effect of Neochetinaeichhorniae for biological control of waterhyacinth. In:Freeman, T.E., ed., Proceedings of the 4th InternationalConference on the Biological Control of Weeds,Gainesville, Florida, 87–92.
Pettet, A. 1964. Seedlings of Eichhornia crassipes: apossible complication to control measures in the Sudan.Nature, London, 201, 516–517.
Reddy, K.R. and Debusk, W.F. 1984. Growth characteristicsof aquatic macrophytes cultured in nutrient-enrichedwater: I. Water hyacinth, water lettuce, and pennywort.Economic Botany, 38, 229–239.
Romero, L.F. 1989. Modelos matemáticos de crecimientodel lirio y políticas de manejo. En: Control y aprove-chamiento del lirio acúatico en Mexico. ComisiónNacional del Agua, Instituto Mexicano de Tecnologia delAgua. Serie de divulgación, 17, 51–78.
Sastroutomo, S., Ikusima, I. and Numata, M. 1978.Ecological studies of water hyacinth (Eichhorniacrassipes (Mart.) Solins) with especial emphasis on theirgrowth. Japanese Journal of Ecology, 28, 191–197.
Sato, H. and Kondo, T. 1983. Biomass production of waterhyacinth and its ability to remove inorganic minerals fromwater. II. Further experiments on the relation betweengrowth and concentration of culture solution. JapaneseJournal of Ecology, 33, 37–46.
Sculthorpe, C.D. 1967. The biology of aquatic vascularplants. London, UK, Edward Arnold, 610p.
Singh, H.D., Nag, B., Sarma, A.K. and Baruah, J.N. 1984.Nutrient control of water hyacinth growth andproductivity. In: Thyagarajan , G., ed., Proceedings of theInternational Conference on Water Hyacinth, UNEP,Nairobi, 243–263.
Trivedy, R.K. 1980. Studies on primary production inshallow, freshwater around Jaipur with reference to theeffects of pollution from domestic sewage. Ph.D. Thesis,Rajasthan University, Jaipur, India.
Vollenweider, RA. 1983. Eutrophication. Notas del segundoencuentro regional de eutroficación en lagos cálidos,Brasil, 16p.
Westlake, D.F. 1963. Comparisons of plant productivity.Biological Reviews, 38, 385–425.
Wooten, J.W. and Dodd, J.D. 1976. Growth of waterhyacinth in treated sewage effluent. Economic Botany,30, 29–37.
Yount, J.L. and Crossman, R.A. Jr 1970. Eutrophicationcontrol by plant harvesting. Journal of the WaterPollution Control Federation, 42, R173–R183.
Zar, J.H. 1974. Biostatistical analysis. USA, Prentice Hall,620p.
120
Water Hyacinth Control through Integrated Weed Management Strategies in Tanzania
G. Mallya, P. Mjema and J. Ndunguru*
Abstract
Integrated weed management (IWM) strategies are having a significant impact on water hyacinth control inTanzania. Water hyacinth has been reduced by over 70% within a period of 3 years. This has been achievedmainly through biological control, manual removal, quarantine regulations, and management of nutrientenrichment. Through manual removal, 60 landing beaches in Lake Victoria were kept free of water hyacinth.Through biological control, two weevils, Neochetina eichhorniae and N. bruchi, have established with adultpopulations of up to 30 per plant. There has been a significant reduction in water hyacinth plant populationdensity, from 45 to 7 plants per 0.5 m2, and large reductions in surface area covered and biomass. Maintenanceand construction of wetlands have been used to minimise nutrient loading in lakes, ponds, rivers and satellitelakes. The management of water hyacinth in rivers and ponds that are acting as potential sources of infestationhas recently begun.
WATER hyacinth is considered to be the most seriousaquatic weed in Tanzania. This free-floating plant ofSouth American origin (Bennet 1967; Jayanth 1988)was observed for the first time in Tanzania in 1955 inthe River Sigi and 1959 in the Pangani River. In 1955it was gazetted as a noxious weed. In recent years,water hyacinth has spread faster, and the most seriousinfestation is in Lake Victoria (Labrada 1995). In1995, about 700 ha of the shoreline including bays andgulfs were affected and by 1998 the coverage was esti-mated at 2000 ha (LVEMP 1999). Water hyacinth hasposed serious environmental and socioeconomic prob-lems in the use and management of water resources.
To mitigate the water hyacinth problem in Tanzania,and in Lake Victoria in particular, integrated weedmanagement (IWM) strategies with emphasis on abiological control program were initiated in 1995under the Lake Victoria Environment ManagementProject (LVEMP), which is a comprehensive,
regional-level environmental program. Under thisproject, each of the riparian countries of Kenya,Uganda, and Tanzania carries out water hyacinthcontrol in the Lake Victoria within its boundaries. Thispaper discusses the use of IWM strategies to controlwater hyacinth in Tanzania, focusing on successachieved so far.
IWM Strategies
Tanzania has integrated biological control, manualremoval of water hyacinth at strategic sites in collabo-ration with local communities, quarantine regulationsand management of nutrient influx into rivers, pondsand lakes to attain sustainable management of waterhyacinth.
Biological control
Work towards biological control of water hyacinthin Tanzania was started in May 1995 when 418 waterhyacinth weevils (Neochetina eichhorniae and N.bruchi) were imported from the IITA Biological
* Plant Protection Division, 1484, Mwanza, Tanzania.Email: [email protected]
121
Control Centre for Africa, Benin into Tanzania andmass-reared at Kibaha National Biological ControlCentre. By June 1995, the weevils had multiplied toover 2000 insects. Between July 1995 and June 1996,9000 adult weevils were released into the Sigi andPangani rivers. Neochetina weevils were released intoLake Victoria for the first time during 1997 and werefollowed by subsequent releases covering all theweed-infested bays, gulfs, ponds and satellite lakes.Eleven (8 community and 3 Institute managed)weevil-rearing units were built around Lake Victoriaeach with 20 to 50 plastic tanks with 500 L capacity.Adult weevils were held on plants to lay eggs. Adultswere then removed and the plants with eggs placedinto the field. Assessment of spread and impact of theweevils is done regularly at both release and recoverysites.
Physical control of water hyacinth
Physical control involves manual removal of theweed using simple tools and equipment. This is aimedat keeping landing beaches, water sources, pumps, andrecreational areas free from water hyacinth. The localcommunities and non-government organisations areconstantly involved in identifying and clearinginfested sites. Hand tools and protective gear worth14.5 million Tanzanian shillings (ca US$20,000) havebeen provided to the community by the government toenhance manual removal work.
Quarantine regulations
To prevent the spread of water hyacinth to weed-free areas, legislation is also being used in Tanzania. Adraft of ‘Water Hyacinth Control Regulations’ wasprepared in September 1999 based on National PlantProtection Act (No. 13 of 1997).
Control of nutrient enrichment
The nutrient conditions and the tropical environ-ment provide fertile conditions conducive to rapidgrowth and proliferation of water hyacinth. Thisproject identifies the different types of nutrients andother environmental factors that promote water hya-cinth proliferation in lakes, ponds, and rivers.
Results
A generally dramatic success in combating water hya-cinth infestation has been realised so far in Tanzania,particularly in Lake Victoria. Water hyacinth is no
longer a menace in the Lake Victoria Basin. Biologicalcontrol has worked very well.
• The water hyacinth infestation in the Lake Victoriahas been reduced by over 70% over a period of 3years.
• Data from ground survey in Lake Victoria hasrevealed only localised water hyacinth infestationsand most of the landing beaches are weed-free.
• Plants estimated to contain approximately 30million eggs of the weevils were placed in the field.
• Preliminary weevil impact data (7 months) fromstabilised water hyacinth mats have revealed asignificant reduction in plant population from 45 to7 plants/0.5m2 (Fig 1).
• Through manual removal, more than 60 landingbeaches in Lake Victoria are kept water hyacinth-free.
• The reproductive index of water hyacinth in LakeVictoria has fallen from 6 to an average 0.5 rametsper plant (Fig 2).
• Both N. eichhorniae and N. bruchi have equallyestablished in release and recovery sites (adultpopulations of up to 30 per plant) but the efficiencyof each species has yet to be determined.
• A survey has revealed the existence of eight waterhyacinth infested ponds in the Lake Victoria basinwith coverage ranging from less than 1 ha to 35.5 haand weevils have been released in some of them.
• Nitrogen and phosphorus have been identified asimportant nutrients in Lake Victoria. They comemainly from industrial, domestic and agriculturaleffluents.
0
5
10
15
20
25
30
35
40
45
50
Mea
n p
lan
t p
op
ula
tio
n p
er 0
.5 m
2
Dec1999
Jan2000
Feb2000
Mar2000
Apr2000
May2000
June2000
Month
Figure 1. Changes in plant population
122
• Wetlands have been constructed and maintained todecrease nutrient loading into the lake.
• Ecological succession is evident in water hyacinthmanagement sites whereby pure water hyacinth isinvaded mainly by water sedges (Cyperus sp.).
Conclusions
The benefits of IWM strategies in dealing with waterhyacinth have been demonstrated in Tanzania. Waterhyacinth in Lake Victoria has been tremendouslyreduced. However, there is a continuous inflow ofwater hyacinth (0.2 to 0.8 ha/day) into Lake Victoriafrom the Kagera and Mara rivers. Furthermore, resur-gence of water hyacinth has been observed in someparts of the water hyacinth managed areas, mainlyfrom seed reserves, and there is a pressing need tomanage them. The continuous presence of water hya-cinth in the lake and the conditions that supported itsrapid spread are still in place. Future work to preparefor any renewed infestation should include:
• research on the relationship between Neochetinaweevils and water hyacinth;
• examining the possibility of using other waterhyacinth control methods;
• carrying out socioeconomic impact assessment ofthe water hyacinth management strategies inTanzania; and
• development of a surveillance system that wouldprovide timely information regarding the location,relative size and rate of increase of water hyacinthmats along the shoreline. A network of observers(mainly fishermen) will be developed to report anycharges in hyacinth coverage within the area.
Acknowledgment
This work is funded by the World Bank and GlobalEnvironmental Facility in collaboration with theGovernment of Tanzania.
References
Bennett, F.D. 1967. Notes on the biological control of waterhyacinth Eichhornia crassipes. PANS (C) 13, 304–309.
Jayanth, K.P. 1988. Successful biological control of waterhyacinth (Eichhornia crassipes) by Neochetinaeichhorniae (Coleoptera: Curculionidae) in Bangalore,India. Tropical Pest Management, 34, 263–266.
Labrada, R. 1995. Status of water hyacinth in developingcountries In: Charudattan, R., Labrada, R., and Center,T.D., ed., Strategies for water hyacinth control. Report ofa panel of experts meeting, 11–14 September 1995, FortLauderdale, Florida, USA. Rome, FAO.
LVEMP (Lake Victoria Environment Management Project)1999. Review of progress on implementation of waterhyacinth control (July 1997–June 1999). Paper presentedat a regional workshop on LVEMP implementation,Mwanza, 1–5 November 1999, 1–15
0
1
2
3
4
5
7
6
Mea
n n
um
ber
of r
amet
s/p
lan
t
Dec1999
Jan2000
Feb2000
Mar2000
Apr2000
May2000
June2000
Month
Figure 2. Changes in the number of ramets per plant
123
Integrated Control of Water Hyacinth on the Nseleni/Mposa Rivers and Lake Nsezi,
Kwa Zulu-Natal, South Africa
R.W. Jones*
Abstract
Water hyacinth infestations on the Nseleni/Mposa River system were sprayed with a herbicide on an ad hocbasis between 1983 and 1995, with no real results being achieved. During the summer of 1985–86, the firstbiological control agent, the weevil Neochetina eichhorniae, was introduced into the system, and by the end of1986 beetle activity, estimated by adult feeding scars, was common throughout. During 1995, a formalIntegrated Water Hyacinth Control Programme was introduced to form a holistic approach to use the variouscontrol options that were available; i.e. chemical, mechanical and biological.
A committee comprising all parties/communities adjacent to the rivers and lake that were affected by thewater hyacinth was formed to monitor the new integrated control program. The program consists of four maincomponents, namely: Survey, Plan, Control and Record. The Nseleni River (17.1 km affected), Mposa River(4.9 km affected) and Lake Nsezi (≈ 260 ha) have been divided into eight management units.
By using the integrated control approach, a total of 18.9 km of river has been cleared of water hyacinthbetween 1995 and the present. The management units that have been cleared of water hyacinth, now requireonly occasional follow-ups to spray any regrowth with a herbicide or to physically remove it. Recent recordsindicate that previously recorded ‘red data’ species of avifauna have returned to the area, namely bitterns(vulnerable and rare), storks (rare) and African finfoot (indeterminate). Oral reports from the local ruralcommunities that rely on fish as a source of food, indicate that their catches have improved—a sure sign thatthe control of water hyacinth in the system is having a positive ecological impact.
Also of importance is the fact that there is reduced evapotranspiration because of removal of water hyacinth,which in turn makes more water available to the environment, industry and the surrounding communities, bothrural and urban. The rural communities have benefited directly, as they are now able to fish and thereby feedtheir families. As a result of the success of this control program, the entire catchments of the Mposa River(Mbabe and Nyokaneni rivers) have been included in the program.
WATER hyacinth was first recorded in South Africa(Cape Province and Kwa Zulu-Natal) in 1910 (Gopal1987). It is believed to have been introduced as anornamental aquatic plant and has since been spread tonumerous localities throughout the country by gar-
deners, aquarium owners and boat enthusiasts (JacotGuillarmod 1979). The main distribution occurs fromlow-lying subtropical to high elevations where frostoccurs (Cilliers 1991).
Water hyacinth is not the only problematic alienaquatic plant in South Africa, as other aquatic plantssuch as parrot’s feather (Myriophyllum aquaticum),red water fern (Azolla filiculoides), water fern (Sal-vinia molesta), water lettuce (Pistia stratiotes), the
* Kwa Zulu-Natal Nature Conservation Service, PO Box10416, Meerensee 3901, Richards Bay, Republic of SouthAfrica. Email: [email protected]
124
reed (Arundo donax) and the bullrush (Thypha cap-ensis) also occur. Water hyacinth it is believed to bethe most problematic. In addition, plants such as thetwo reeds Phragmites mauritianus and P. australishave been identified as being plants with future majorimpact possibilities (C.J. Cilliers, pers. comm.).
Water hyacinth is a declared weed in South Africaand is covered by legislation. This is the Conservationof Agricultural Resources Act (Act 43 of 1983) and isadministered by the Directorate of Resources Conser-vation of the National Department of Agriculture. TheAct states clearly that this weed must be controlled.The South African Department of Water Affairs andForestry (DWAF) is mandated to co-ordinate thecontrol of water hyacinth and to execute measures insituations where the weed threatens state water works.In other scenarios it becomes the responsibility of theprovincial and local water authorities.
Water hyacinth was first recorded on the Nseleni/Mposa Rivers (28°42'S; 32°02'E) and Lake Nsezisystem in 1982 (Ashton 1982) and is believed to havebeen introduced in approximately 1978 as an orna-mental plant. At this stage the water hyacinth infesta-tion had already been recognised as being a problem,covering an area of approximately 1.5 km2. Therewere serious concerns that damage would be causedto: (1) the national road bridge over the Nseleni River,and (2) the functioning of the two water treatmentfacilities on Lake Nsezi.
The Nseleni and Mposa rivers, as well as LakeNsezi, are used by the surrounding rural communitiesto supplement their daily food with fish catches. Formany, the fish they catch are their main source ofdietary protein. Both large-scale sugarcane farms andsmall-scale subsistence farms also irrigate from theriver. Mhlathuze Water Board (the local waterauthority) pumps water from Lake Nsezi and supplieswater for both domestic and industrial use to thegreater Richards Bay and Empangeni areas. In addi-tion, extraction points for Richards Bay Minerals(mining) and the rural town of Nseleni are located onthe Nseleni River. A sewage plant that serves the ruraltown of Nseleni is located on the bank of the MposaRiver.
The Kwa Zulu-Natal Nature Conservation Service(KZNNCS) at one stage offered boat trips for birdviewing on the Nseleni River. These were abandonedas water hyacinth encroached on the river. Rarespecies of avifauna, like the African finfoot (Podicasenegalensis) and other aquatic fauna and flora disap-peared from the area, as a result of the increase in thewater hyacinth infestation. In addition, the rural com-
munities were not only unable to fish, but also found itimpossible to cross the Nseleni River to get to theirwork places on farms.
Control Efforts
Ad hoc control efforts were practised between the late1970s and 1994 by various interested and affected par-ties. By 1982, stretches of the Nseleni and Mposarivers were covered with water hyacinth (100% cov-erage) and KZNNCS initiated control of the weed. In1984, a heavy flood alleviated the problem, as most ofthe water hyacinth was washed away before an aerialspraying operation could be implemented. Thereafter,little was done to the remaining islands of water hya-cinth, because the decreased level of infestation wasno longer seen as a threat.
Chemical control was reintroduced in the mid 1980swhen the Nseleni River was once again covered bywater hyacinth, but there was no management plan andchemical spraying was carried out on an ad hoc basis.It is important to note that eradication/control stepswere undertaken only once the water hyacinth becamea problem.
In an independent effort, the Plant ProtectionResearch Institute (PPRI) of the Agricultural ResearchCouncil imported (via Australia) a weevil, Neochetinaeichhorniae, releasing 1400 adult insects on theMposa River in December 1985. By November 1989,most of the water hyacinth had once again washedaway as a result of exceptionally heavy floods. How-ever, the biological control agents persisted on theremaining water hyacinth.
To put a monetary value on the economic losscaused by water hyacinth on nearly 22 km of river and360 ha of lake proved to be extremely difficult,because of the ad hoc control efforts that were imple-mented. The cost to KZNNCS in just keeping the riveropen on its 6.3 km of river boundary amounted toR15,0001 in 1991. When the infestation was at itsheight by the mid 1990s, it cost R20,000 to clear sec-tions of the river.
Mhlathuze Water Board remains opposed to andconcerned about any possible large-scale chemicalspraying of water hyacinth and the effect the decayingorganic material would have on the odour and taste ofthe water. In addition, it also feels that water hyacinthpartly purifies the water, because of the nutrients ittakes up. Large-scale aerial spraying could also havedetrimental environmental effects on lake and riparian
1 R = South African Rand. In March 1991, R3.67 = US$1.
125
vegetation such as Papyrus spp. and Barringtoniaracemosa, as well as other indigenous flora. Thiswould be undesirable and ecologically unacceptable.In addition, any uncoordinated large-scale chemicalspraying at the wrong time would nullify the effect ofthe biological control agents, as all sessile stages arekilled when plants are sprayed when there is not a peakof adult insects.
In an additional independent effort, PPRI (Pretoria)introduced Niphograpta albiguttalis (150 moth larvae)and a mite Orthogalumna terebrantis (800 adults) inJanuary 1994, in an attempt to supplement the previ-ously introduced weevil N. eichhorniae. At the sametime as these biological control agents were released,water hyacinth plants were inspected for weevildamage. Results indicated that the weevil N. eichhor-niae had spread throughout the system, which was apositive sign.
During 1994, an aerial survey was undertaken in anattempt to record the extent of the water hyacinthinfestation in the entire system. The results of thesurvey indicated the infestation varied between 100%and approximately 40% coverage in different sectionsof the system.
Integrated Water Hyacinth Control
In March 1995, an Integrated Water Hyacinth ControlCommittee was formed. This committee met regularlyand welcomed other representatives from the commu-nity to attend these meetings. It also held ‘open days’to show the community the results achieved.
The first objective of the committee was to collateall the work that had previously been carried out on thewater hyacinth infestation and to formulate a holisticapproach to use the various control options that wereavailable; i.e. chemical, biological and mechanical. Inaddition, a management plan was formulated, con-sisting of four main components, namely Survey, Plan,Control and Record, as well as an action plan for whenfloods occurred.
A map of the system (affected areas: Nseleni River– 17.1 km, Mposa River – 4.9 km and the Nsezi Lake– 268 ha) was drawn up and used to designate eightmanagement units (MUs) of controllable size (Fig. 1).Further to this, each MU was assigned a level of con-trol, i.e. total control or containment, as well as theappropriate method of control, i.e. chemical, biolog-ical, mechanical or a combination of control methods.
It was emphasised at the outset that the managementplan was a working document and that objectives andcontrol methods would change as work progressed. In
addition, the committee emphasised and recognisedthat total eradication was impossible, because of thelong lived seed source. Water hyacinth seed can liedormant for up to 14 years (Penfound and Earle 1948).It was therefore recorded that total maximum accept-able percentage coverage would be 20%.
Each MU was assigned to an individual, organisa-tion or company. For example, MU 1 was assigned toa sugarcane farmer and KZNNCS, MU 2 to KZNNCS,MUs 3 to 5 to MONDI Forestry and KwaMbonambiConservancy, MUs 6 and 7 to KZNNCS and MU 8 tothe local water authority—the Mhlathuze WaterBoard (to merely inspect and report on the status ofbiological control agents).
In March 1995 it was stated that the objective forMUs 1 to 4 would be total control using all methodsavailable, and that containment of the infestation usingbiological control agents in MUs 5 to 8 would takeplace. Further to this, various sectors from the commu-nity were assigned MUs to control. Awareness cam-paigns were run at the same time, through lectures,radio talks and articles in the local press. Instead ofusing labour from the local rural community to removewater hyacinth manually, school children and theirelders were successfully prompted to replant and sta-bilise the banks of the river with suitable indigenousvegetation where they had previously chopped downtrees to practise subsistence farming. This was donebecause of the threat from crocodiles in the river,which killed several children every year. This wasunrelated to water hyacinth control.
In an attempt to reduce the spread of water hyacinthseed and to make the chemical control cost effective,permanent cable booms (28 mm steel) were placedacross the river at the confluence of the Mposa andNseleni rivers (MU5), at the southern end of MU2 andat the northern end of MU6. Cables were also installedacross the river where MUs 6 and 7 met and whereMUs 7 and 8 met. The cables were placed in such amanner that they hung beneath the surface of thewater, thereby catching the root system of the waterhyacinth. Plastic buoys (donated by the Richards BayCoal Terminal) were used as flotation on the cables.Note that each permanent cable has a ‘weak link’ in it.Previous experience showed that during floods, notonly was there a vast volume of water, but that thecable anchors (trees) were unable to hold the weight ofthe water hyacinth that built up on the cables.
In addition to the permanent cables across the river,temporary cables placed across MUs 1, 2 and 6 toallow the water hyacinth to back-up against them,which assisted the chemical control method.
126
Further assistance to the control program occurredwhen the Nseleni sewerage works on the Mposa Riverwas upgraded, and the effluent quality improved dra-matically. Before upgrading, the ammonia (NH3) was14.2 ppm and the chemical oxygen demand (COD)130 ppm. After commissioning, the ammonia droppedto 1.2 ppm and the COD to 53 ppm, a vast improve-ment. However, it was further recorded that nutrients
were entering the system from adjacent sugarcanefarms and forestry areas.
During the course of 1995, a total of approximately2400 litres of glyphosate had been sprayed in MUs 1, 2,3, 5 and 6 and seven river patrols were carried out tomonitor water hyacinth infestations, inspect the effectthat spraying and biological control had on water hya-cinth and to carry out routine maintenance of the cables
Railway line
N2 (to Empangeni)
Nseleni River
MU 1(6700)
EnseleniNature
Reserve
EnseleniNature
Reserve
MU 2(1700)
MU 6(4600)
MU 5(1200)
Nseleni Village
MU 4(2300)
MU 3(1400)
Mpo
sa Rive r
N2
(to
KwaM
bona
mbi
)
MU 7(2900)
to Richards Bay
0 1 2 3 4 5
N
Kilometres
MU 8(212 ha)
(29 ha)
(27 ha)
Figure 1. Management units (with length of units in metres) for water hyacinth control on theNselini/Mposa rivers and Lake Nsezi
127
(total cost R9345.00). As a result of the success levelsachieved by the end of 1995, the control committeeagreed to adjust the management plan objectives and toelevate MU6 to total control and to retain MUs 7 and 8as containment MUs (biological control only).
Because the local water authority needs to removewater hyacinth from its inlet screens, it has agreed toremove biological control agents from the water hya-cinth and to return them in Lake Nsezi, therebyensuring that they are maintained in the system.
In March 1996, the first release (50 adults and 100nymphs) was made of a new biological control agent,the mirid Eccritotarsus catarinensis, at the entrance toLake Nsezi. During June 1996, a further 500 adult E.catarinensis were released. It was reported at the June1996 meeting that no chemical spraying had been donein MUs 1, 2 and 6. Because of the decreased infesta-tion level of water hyacinth, these units had merelybeen monitored. The status of biological controlagents throughout the system was positive, with one ormore agents being recorded in the MUs where waterhyacinth infestations occurred. In addition, the path-ogen Cercospora piaropi was found on some plants,and the fungus Acremonium zonatum was recorded forthe first time.
During October 1996, another weevil species, Neo-chetina bruchi, was obtained from PPRI (Pretoria), aswell as additional E. catarinensis (10 infested plants),and these were released into the system. In addition,the management plan objectives were again adjustedto reflect the progress being made. The managementplan now allowed for total control in MUs 1–7, withonly MU8 designated for containment (biologicalcontrol only). It was also agreed to drop the totalallowable coverage percentage from 20% to 10%.
Records indicate that financial expenditure oncontrol of water hyacinth during 1996 fell toR5892.00.
During 1997, glyphosate herbicide continued to beapplied to water hyacinth in MU 5 (100% infestation)and MU 7 (infestation increased to 60%), withvarying amounts of success. It is important to realisethat the islands of water hyacinth are left after theapplication of chemicals. This is to allow the biolog-ical control agents to continue to move within thesystem.
Entry into the Mposa River (MU 5) from the southbecame extremely problematic, because not only wasthere a 100% infestation of water hyacinth, but alsoindigenous aquatic vegetation had severely encroachedon the area. Of note was the invasion of Echinochloapyramidalis, an indigenous perennial plant, which
enjoys moist terrestrial or aquatic conditions and useswater hyacinth as a substrate on which to form densestands. Other possible contributing factors towards theestablishment of E. pyramidalis, are nutrient enrich-ment of water and silt-laden watercourses.
During 1997, a distance of approximately 200 mwas gained into MU 5, from the southern side. In addi-tion, MU 6 had to receive attention, because the totalallowable percentage coverage exceeded 10%. Some296 litres of glyphosate was used in MUs 5 and 6, withthe required result being achieved. The status of bio-logical control agents in MUs 3, 4, 5, 7 and 8 remainedpositive. A further 300 adult E. catarinensis werereleased on Lake Nsezi during the latter part of 1997.
By September of 1998, a further 238 litres ofglyphosate had been applied to MUs 5, 6 and 7. Some28 hours and 78 labour units, over a period of 25 days,were expended to inspect, carry out cable maintenanceand chemically spray the water hyacinth infestations.The results of the water hyacinth infestation inspec-tions indicated a high percentage of biological controlagent activity throughout the system. During August1998 a setback occurred when an area of approxi-mately two hectares of water hyacinth was blown fromMU 7 into MU 6, during a period of exceptionallystrong southeasterly winds. Fortunately, the cable didnot break and a high percentage of water hyacinthremained in MU 7. With the aid of temporary cablesthe approximately two hectares of water hyacinth thathad blown into MU 6 and which had subsequentlybroken up into smaller pockets, was cordoned off andchemically sprayed.
A major injection to the control program in 1998was the assistance received from the MONDI forestscompany, which achieved excellent chemical controlresults in MU 4 (Mposa River). Between May andOctober 1998, MONDI spent R2800 per month onchemicals and labour, to open up stretches of theMposa River from both water hyacinth and invasiveindigenous aquatic plants. In addition, KwaMbonambiConservancy approached various industries in Rich-ards Bay in an effort to get them to become involved inthe project. The result of this drive was thatR38,000.00 was received (to purchase new sprayingequipment and an outboard engine) and MONDI Kraftoffered to construct a barge-like boat and a trailer(approximately R50,000) which would be used inspraying. Further to this, KwaMbonambi Conservancypledged 200 litres of roundup and Richards Bay Min-erals pledged R6000 towards the project.
A flood during February 1999 opened up about 1 kmof MU 5, and MUs 1, 2, 6 and 7 became 98% free of
128
water hyacinth. The infestation in Lake Nsezi haddropped dramatically to approximately 35%. DuringMay 1999, for the first time in many years, members ofthe control committee were able to proceed from alaunch site in MU 2 and travel all the way to theMhlathuze Water Board extraction point on the south-east bank of Lake Nsezi (MU 8). Biological controlagents persisted on the remaining water hyacinth.
As a result of the high success rate achieved withthe integrated control on the Nseleni River and a smallsection of the Mposa River (MU5), it has beendecided to expand this project to include the catch-ment of the Mposa Rivers, namely the Mbabe and
Nyokaneni rivers. A management plan is currentlybeing drawn up to focus on 14 management units onthese rivers (Fig. 2).
Community Involvement
Community involvement has no doubt been the secretof the success of the integrated control program.Although the control of water hyacinth was initiatedby staff from the Enseleni Nature Reserve (KZNNCS),it soon became apparent that additional assistancewould be required from the surrounding community,as well as the ‘end users’ of water, i.e. industry and
MU 11
MU 9
MU 9
MU12
MU13
MU 14MU 10
MU10
Mposa
MU 8MU7
MU 6
MU 3
MU 4
MU 5 MU 2MU 1
Nyokanen i
Mb abe
Mapelane
Sappi/M
ondi boundar
y
NseleniTownship
Kwambo
Railwayline
N2
0 1 2 3 4 5
N
Kilometres
Figure 2. Water hyacinth management units of the Mbabe and Nyokaneni rivers
129
urban communities. The surrounding communitiesbecame involved in the project because they dependon the water resource directly for their livelihood(fishing and agriculture), or as an extractable resource(for mining, industrial and urban uses).
It is important to note that, although such projectsrequire vast amounts of funding in the initial stages tobring the infestation under control, a level will bereached where only maintenance will be required andtherefore a set annual funding requirement must beobtained. However, funding requirements willdiminish only if there is enthusiasm, success and astable authority responsible for the implementation ofthe project.
Conclusion
By using an integrated control approach, between1995 and the present, a total of nearly 22 km of riverhas been cleared of the original infestation of waterhyacinth. The sections that have been cleared of waterhyacinth now require only occasional follow-up toremove any regrowth. Recent records indicate thatpreviously recorded ‘red data’ species of avifaunahave returned to the waterways. Reports from the ruralcommunity, which relies on fish as a source of food,indicate that their catches have improved—a sure signthat the clearance of water hyacinth in the system isproducing a positive ecological impact.
The advantages of controlling water hyacinth infes-tations far outnumber the disadvantages.
Water, as a natural resource, is for many reasons fastbecoming a dwindling resource, and thereforedemands especial attention.
Because of the success achieved with the integratedcontrol program, the entire Mposa River catchment,
i.e. the Mbabe and Nyokaneni rivers, has now beenincluded in the control program.
Uncoordinated efforts to control water hyacinth onthe same system by different parties have proven to bea waste of time and money. Once a proper integratedmanagement plan and control is implemented, waterhyacinth infestations can be reined in. Nevertheless,prevention is better than cure, and it is of the utmostimportance that infestations of water hyacinth be con-trolled before they become a problem.
The Nseleni/Mposa rivers and Lake Nsezi scenariois an example of what can be achieved on limitedbudgets but with vast amounts of enthusiasm.
Acknowledgments
For their contributions to the program, I thank PPRI(Pretoria), Enseleni and KwaMbonambi Nature Con-servancies, MONDI and Sappi Forests, Mondi Kraft,Mhlathuze Water Board, Richards Bay Coal Terminal,Richards Bay Minerals, the rural community ofNseleni township and the KZNNCS staff at the Ense-leni Nature Reserve.
References
Ashton, P.J. 1982. Report to Empangeni Municipality—avisit to the Nseleni River and Lake Nsezi to the waterhyacinth problem. Council for Scientific and IndustrialResearch. National Council for Water Research.
Cilliers, C.J. 1991. Biological control of water hyacinth,Eichhornia crassipes (Pontederiaceae), in South Africa.Agriculture, Ecosystems and Environment, 37, 207–217.
Gopal, B.1987. Water hyacinth. Elsevier, New York. 471pp.
Jacot Guillarmod, A. 1979. Water weeds in southern Africa.Aquatic Botany, 6, 377–391.
Penfound, W.T. and Earle, T.T. 1948. The biology of thewater hyacinth. PANS, 20(3), 304–314.
130
Preliminary Assessment of the Social, Economic and Environmental Impacts of Water Hyacinth in the
Lake Victoria Basin and the Status of Control
A.M. Mailu*
Abstract
The paper presents preliminary data collected in an assessment of the social, economic and environmental impactsof water hyacinth in the Lake Victoria Basin. A summary of the status of control and strategies for the future isgiven. The report draws on field observations made, studies through interviews of affected communities andorganisations, personal communications and published reports by scientists in the region.
Lake Victoria, the world’s second largest freshwater body, supports an estimated 25 million people living in theBasin, with an estimated gross economic product of US$3–4 billion annually, mainly from subsistence agricultureand fishing in Kenya, Uganda, Tanzania, and parts of Rwanda and Burundi.
The multiple activities in the Lake Victoria Basin have increasingly come into conflict, thus making the lakeenvironmentally unstable and increasingly inviting environmental threats, including infestation by water hyacinth,which has brought social, economic and environmental problems to the communities living in the Lake Basin sinceits first appearance in the Lake in the late 1980s and early 1990s.
The maximum water hyacinth cover in Lake Victoria was reached between 1994 and 1995 when 80% of theshoreline in Uganda was covered with about 4000 ha of water hyacinth, there was about 6000 ha coverage in Kenyaand about 2000 ha in Tanzania. In Rwanda and Burundi, tributaries feeding into River Kagera currently continueto discharge mats of water hyacinth into the lake at about 3.5 ha per day. The status as at June 2000 was slightlydifferent, with scant water hyacinth in the Uganda side of the lake and much disintegrated and stunted waterhyacinth in Kenya and Tanzania sides of the lake and the scene is now dominated by hippo grass.
Impact assessments of water hyacinth have generally been subjective, with few quantitative outputs. However,over the last nine years or so, water hyacinth has had a negative impact on the organisations and communities inthe Basin. Surveys have revealed negative social impacts including lack of clean water, increase in vector-bornediseases, migration of communities, social conflict and difficulty in accessing water points. Important economicimpacts readily perceived by Basin communities have included reduced fish catches, increase in transportationcosts, difficulties in electricity generation and water extraction, fewer tourists, blockage of irrigation canals andenvironmental impacts such as decline in water quality, water loss through evapotranspiration, siltation, increasedpotential for flooding and a decline in the diversity of aquatic life.
Although water hyacinth has posed serious economic, social and environmental consequences, there is hope thatthe control strategies already adopted will continue to reduce deleterious impacts and allow sustained developmentin the Lake Victoria Basin. There is, however, a great need to undertake research to quantify the levels of damage,and the costs of control, loss of livelihood, disease, and disruption of normal operations caused by water hyacinth.
* Kenya Agricultural Research Institute, PO Box 57811,Nairobi, Kenya. Email: [email protected]
131
LAKE Victoria, with a surface area of 68,800 km2 andan adjoining catchment of 184,000 km2, is the world’ssecond largest body of fresh water, second only toLake Superior. Lake Victoria touches the equator in itsnorthern reaches, and is relatively shallow. Itsmaximum depth is about 80 m and the average about40 m. The lake’s shoreline is long (about 3500 km) andconvoluted, enclosing innumerable small, shallowbays and inlets, many of which include swamps andwetlands which differ a great deal from one anotherand from the off-shore environment of the lake.
Kenya, Tanzania and Uganda control 6, 49, 45%,respectively, of the lake surface. The gross annual eco-nomic product from the lake catchment is in the orderof US$3–4 billion, and it supports an estimated popu-lation of 25 million at per capita annual incomes in therange US$90–270. The lake catchment thus providesfor the livelihood of about one third of the combinedpopulations of the three countries, and about the sameproportion of the combined gross domestic product.The lake catchment economy is principally an agricul-tural one, with a number of crops (including exports offish) and a high level of subsistence fishing and agri-culture. In Kenya and Uganda, the areas of coffee andtea in the catchment are a significant part of thosenations’ major agricultural exports. The quality of thephysical environment is therefore a fundamental factorin maintaining and increasing the living standards ofthe growing populations in Kenya, Uganda, Tanzania,Rwanda and Burundi (Labrada 1996).
Major Threats to Lake Victoria
The lake is used as a source of food, energy, drinkingand irrigation water, transport, and as a repository forhuman, agricultural and industrial waste. With the pop-ulations of the riparian communities growing at ratesamong the highest in the world, the multiple activities inthe Lake Victoria Basin have increasingly come intoconflict. This has contributed to rendering the lake envi-ronmentally unstable. The lake ecosystem has under-gone substantial, to some observers alarming, changes,which have accelerated and are increasingly dominatedby the potentially toxic blue–green algae (Twongo andBalirwa 1995). The frequency of water-borne diseaseshas increased. Water hyacinth, absent until as recentlyas 1989, has choked important waterways and landings.Overfishing and oxygen depletion in the deeper watersof the lake threaten the artisanal fisheries and biodiver-sity (over 200 indigenous species are said to be facingpossible extinction). Scientists advance two mainhypothesis for these extensive changes. First, the intro-
duction of Nile Perch as an exotic species some 30 yearsago has altered the food web structure. The secondhypothesis is that nutrient inputs from its catchment arecausing eutrophication. Thus, although the lake and itsfishery have shown evidence of dramatic changes in thepast three decades, the problems have arisen mainly asa result of human activities in the lake basin (Bugaari etal. 1998; Freilink 1991; Goodland 1995).
Aquatic Weeds in East Africa
The role of plants as a vital component of the aquaticenvironment is recognised as that of provision of suit-able shelter and food for other plants and animals.Some plants, especially when transported to non-endemic areas, will undergo rapid growth and repro-duction to out-compete native plants, becoming trou-blesome. They may thus become weeds that causeenvironmental and socioeconomic problems. Threeplants, namely water hyacinth, salvinia which is alsoknown as Kariba weed or water fern (Salvinia molesta(Salviniaceae)) and water lettuce (Pistia stratiotes(Araceae)) cause the most serious problems in sub-tropical regions. They are also the main freshwateraquatic weeds in the Lake Victoria Basin. Other weedsthat are of minimal importance are azolla (Azolla fil-liculoides), parrot’s feather (Myriophyllum aquatica)and pennywort (Hydrocotyle ranunculoides).
Aquatic weeds, including water hyacinth, grow bestin tropical and subtropical climates where temperaturesin the range 25–30°C favour their growth almost allyear round. The Lake Victoria Basin is one such areawhere sufficient nutrients are available and have thusadded to the rapid growth of the weed (Mailu 1998a).
Water Hyacinth in East Africa
Although water hyacinth is known to have been kept inNairobi and Mombasa as an ornamental plant sinceearly 1957, it appeared in the natural water systems ofEast Africa in Tanzania (River Sigi) in 1956 and laterin the Pangani River. In Uganda and Kenya, reportsindicate that the weed appeared in Lake Kyoga(Uganda) and Lake Naivasha (Kenya) and later inLake Victoria in the late 1980s (Twongo 1993, 1996;Twongo and Balirwa, 1995; Taylor 1993). It was,however, between 1990 and 1992 that the negativeimpacts of the weed started becoming evident alongthe shores of the three riparian states. The weed hascontinued to impact negatively on the lives of the lake-side communities living in the Lake Victoria Basin,especially those closest to the lakeside.
132
Magnitude and Distribution
Weed distribution and cover in Uganda
Preliminary monitoring of the stationary fringe ofwater hyacinth in Lake Kyoga demonstrated that themaximum cover was reached in 1994, when about 570ha of water hyacinth was distributed along close to60% of the shore. Maximum cover of stationary matsalong the banks of River Nile was also estimated in1994 to fringe about 80% of shoreline length, withtotal water hyacinth cover estimated at about 500 ha.In Lake Albert in Uganda, the weed remained confinedto the northern and southern extremities of the lake,probably because of the turbulence of the lake and theabsence of extensive sheltered, shallow banks alongother shores, where water hyacinth could anchor. Inthe Uganda portion of Lake Victoria, the stationaryfringe of the water weed stabilised in 1995 when itsstretched along about 80% per cent of the shorelineand covered an estimated 2200 ha. Maximum cover ofthe mobile components of the weed in this lake cameto about 1800 ha in 1998 (Twongo et al. 1995). Hence,the maximum cover estimate for water hyacinth in theUganda portion of Lake Victoria was 4000 ha. Esti-mates made in April 1999 and in August 1999 indi-cated that the input of water hyacinth into LakeVictoria through the River Kagera was 3.5 ha perweek. However, it was noted that a significant quantityof this influx of water hyacinth into Lake Victoria bythe river was fragmented in the vicinity of the rivermouth by wave action. Similarly, considerable quanti-ties of the mobile mats of the water weed were alsodestroyed during its annual long distance movementaround the Lake Victoria, including by rainstorms andprevailing winds. Currently only small remnants ofwater hyacinth plants are to be found in Ugandanwaters of Lake Victoria, except in the region whereKagera River enters the lake.
Weed distribution and cover in Kenya
In the Kenyan part of Lake Victoria, the infestationof water hyacinth oscillates during the year amongstthe following bays: Kisumu, Kendu, Nyakach, Homaand Asembo. Some water weeds occur also inWichlum and Uharia in Siaya; and Rukapa andMabinyu in the Nzoia River Delta in Busia District.Infestations are insignificant in Migori and Suba dis-tricts. In April 1998, estimated cover of water hyacinthwas 1000 ha in Kisumu Bay, 3200 ha in Nyakach Bay,600 ha in the Sondu Miriu Delta and 1200 ha in Osodo
Bay. Total cover in the Kenyan waters of Lake Vic-toria is estimated at 6000 ha. However, a survey con-ducted in mid August 1999 indicated a decrease in thecover in Kisumu, Kendu and Homa bays. The indica-tions are of increasing disintegration of the weedmasses, to a point where, in certain areas, ecologicalsuccession replaces the original weed mats. A majorcontribution to the control of stationary mats of waterhyacinth along lake shores and river banks in EastAfrica has been the result of ecological succession, theprogressive displacement of one or more species ofplants (and animals) by another. Pure mats of thewaterweed were invaded initially by aquatic ferns and/or sedges, often to be followed progressively by hip-pograss (Vossia cuspidator) which invariably eventu-ally dominated and shaded out the water hyacinth.Ecological succession has made a major contributionto the control of fringing water hyacinth in LakeKyoga and in the Ugandan portion of Lake Victoria.As at April and August 1999, stunted water hyacinth indisintegrated mats and weed succession were clearlyevident in most parts in Kenyan side of infested partsof the Lake (Mailu 1998b).
Weed distribution and cover in Tanzania
Surveys conducted in April and mid August 1999indicated that water hyacinth infestations in theportion of Lake Victoria in Tanzania were located inMara Bay, Bauman Gulf, Speke Gulf, Mwanza Gulf,Emin Pasha Gulf and Rubafu Bay. Currently, waterhyacinth occurs also in the Kagera, Sigi and Panganirivers, as well as in streams and water ponds aroundDar-es-Salaam and close to Lake Victoria. The totalcover estimate of water hyacinth in the Tanzanianwaters of Lake Victoria was 2000 ha. As in Kenya,increased weed stunting and disintegration of originalmats indicated that water hyacinth had experiencedsevere environmental stress, including that occasionedby the weevils already released into the lake.
Weed distribution and cover in Rwanda and Burundi
Currently, verbal reports indicate that water hya-cinth infestations are located in the major rivers thatfeed into the River Kagera from both Rwanda andBurundi. Surveys in April and August 1999 demon-strated that fairly large mats of the weed were floatingand moving into the Kagera River and onward intoLake Victoria at the rate of approximately 3.5 ha per
133
day. The status of water hyacinth infestation in theinland waters of Rwanda and Burundi is not clearlydocumented (Lowe-McConnel et al. 1992).
Impact of Water Hyacinth in the Lake Victoria Basin
The concern over water hyacinth
Water hyacinth infestation has resulted in serioussocioeconomic and environmental problems for mil-lions of people in riparian communities. Normally, theweed proliferates to form extensive floating mats thatcause disruption in electricity generation, irrigationcanals, navigation and fishing activities, and cause anincrease in water loss through evapotranspiration. Theweed also reportedly provides breeding grounds forschistosome (bilharzia)-carrying snails and malaria-carrying mosquitoes. The cost of water hyacinth infes-tation for countries in the region is estimated to be ofthe order of billions of dollars. For example, in LakeVictoria, the infestation currently covers 12,000 haand is affecting the livelihoods of more than 40 millionpeople in Kenya, Tanzania and Uganda. By the end of1997, media agencies reported a 70% decline in eco-nomic activities at the Kenyan port of Kisumu as aresult of the water hyacinth choking the port and fishlanding grounds. The rapid proliferation of water hya-cinth in the region is a result of the absence of naturalenemies, and the widespread availability of nutrientsin freshwater bodies. The nutrient-enrichment offreshwater bodies in the region is a result of pollutionand other factors arising from the rapid increase ofhuman population and corresponding activities inurban and rural areas. Large urban sewers and othereffluent discharges are well known sources of pointwater pollution, while extensive use of improper agri-cultural methods, and land uses that often result in soilerosion, are a major source of non-point water pollu-tion. Soil erosion, especially when caused by water,carries soil nutrients into the water bodies down hill.
Water hyacinth impacts on lakeside communities
Social and economic impactsThe impacts of water hyacinth may be categorised
into social, economic and environmental. Scientists inthe region have attempted to gather secondary data onthe impact of the weed on the lakeside communities onthe Kenyan, Tanzanian and Ugandan sides of the lake.This information was collected through interviews
with officers-in-charge of district hospitals, fisheries,water supplies in municipalities, cargo and humantransportation. Organisations and communities affect-ed by water hyacinth infestation were thus identified inthe various countries. They are listed in Table 1.
An indication relating to the perceptions of a cross-section of the communities and agencies in East Africaon impacts, as well as on control strategies of waterhyacinth, was obtained through a survey, the results ofwhich are recorded in Table 2. Lakeside communitiesdeemed socioeconomic impacts more important thanenvironmental impacts. The real costs and quantifiedimpact levels were, however, not clear to the commu-nities. Most of those interviewed identified decrease infish catches, increase in certain diseases, increasedtransportation costs, and difficulties associated withclean water availability as major negative impacts.
Fish production
Information from the Fisheries Department, Kenyaindicated that there was a 28% increase in total annualfish catches between 1986–1991 and 1991–1997, from133,097 tonnes to 169,890 tonnes. There was anincrease in all species of fish caught except Oreo-chromis, Clarias and Mormyrus, which showeddeclines of 14, 37 and 59%, respectively, over thesame period. These declines may have been associatedwith the inability of fishermen to access the fishinggrounds for those species because of water hyacinthinfestation.
Generally therefore, as a result of water hyacinthinfestation, accessibility to land and water has beenhindered, resulting in reduced fish catches, especiallyof tilapia and mudfish which are found mainly alongthe shores. Fisherfolk, however, reported increasedfish catches from suitable breeding grounds providedby water hyacinth e.g. tilapia, synodontis, protopterusand labeo. There is, however, need to clarify this con-flicting information; in many more areas around thelake. A reduced fish catch would have an adverseeffect on the quality of life of the communities aroundthe lake and consequently affect sustainable develop-ment in the region.
Marine cargo and human transportation
In Kenya, water hyacinth hampered the movementof smaller vessels, especially canoes, used for bothcargo and human transportation. The activities of theKenya Railways have been closed since 1997 at all thepiers in Asembo, Homa Bay, Kendu Bay, Kowor,Mbita and Mfangano, except for Kisumu port, whichis operational for only larger vessels, though they also
134
experience some difficulties. Such vessels (weighingover 700 tonnes) have their propellers deep, avoidingentanglement with water hyacinth. Boats with capaci-ties less than 700 tonnes cannot operate where there isheavy water hyacinth infestation. The Kenya Railwayskeeps 10 people permanently employed to removewater hyacinth from the bridge at the Kisumu pier.Data made available by the Port Officer, Kenya Rail-ways, Kisumu, covering the period 1996–1998,showed an increase in incoming cargo volume (fromabout 43,000 tonnes to 130,000 tonnes), while thereverse was the case for outgoing volume (from about93,000 tonnes to 37,000 tonnes). There were norecords for human transportation and cargo for thesmaller vessels, as the beaches maintained no properrecords. The frequent closure of the Kisumu portaffects the communities in several ways including lossof income and a general decline in sustainable devel-opment in the affected regions.
In Uganda, where observations had been madebetween 1994 and 1997, large vessels were, as inKenya, able to force their way through the weed mats,but physical removal was also necessary in Port Bell.
Vessels required extra time to dock, thus resulting inthe use of more fuel.
Some initial estimates were made for these costs,but additional data and observations were required tovalidate the initial observations. It should be notedthat these figures represent 1995 estimates. At thattime the water hyacinth infestation was increasing ata rapid rate and it was recognised that unless con-trolled, it would spread and become more of aproblem (Twongo and Balwira 1995). In the absenceof a successful control program, the following werethe estimated costs for the five years after 1995 inUganda.
1. Maintaining a clear passage for ships to dock atPort Bell in Uganda: US$3–5 million perannum;
2. Cleaning intake screens at the Owen Fallshydroelectric power plant at Jinja in Uganda:US$1 million per annum;
3. Losses in local fisheries from accumulation ofwater hyacinth at fishing beaches and landingsites around the lake, making it difficult orimpossible for fishing boats to be launched or
Table 1. Organisations and communities affected by water hyacinth
Country Affected organizations/institutions/communities
Uganda Fishing (fishermen, fishmongers, fish processors, consumersRiparian communitiesLake transport (Uganda Railway Corporation)National Water and Sewerage CorporationNational Agricultural Research OrganizationNational Environmental Management AuthorityDepartment of Fisheries
Kenya Fishing communityKenya RailwaysLocal councilsLocal and provincial administrationKenya Agricultural Research Institute (KARI) Fisheries DepartmentKenya Marine Fisheries Research Institute (KEMFRI)
Tanzania Fisheries from fishermen, through traders to scientists Agricultural irrigation Electricity generating Navigation Health Water supply
Rwanda Agriculture; irrigationCommunity developmentVarious non-governmental organisations
Burundi Agriculture; irrigation
135
recovered: US$0.2 million per annum, but witha very serious local impact.
4. Loss of the beaches, water supply for domestic,stock and agricultural purposes: US$0.35million per annum.
Water supply
Water supply to both villages and municipalities isaffected by water hyacinth. In municipalities, waterhyacinth interferes with the water intake pointsthrough blockage, which lowers the quantity of waterpumped. In Kisumu, the municipality reports that thequantity of water supplied has dropped from 20,000m3 to 10,000 m3 per day. Homa Bay reported acapacity of 1400 m3 per day (Table 3) against ademand of 4000 m3 for 50,000 residents. Kisumumunicipality has an alternative water source of 1000m3 per day from the River Kibos; while Homa Baymunicipality has none. This situation thus causes con-stant water shortage in Homa Bay and at certain timesin Kisumu.
Water hyacinth infestations have been reported tolower the water quality in Kenya and Uganda (in terms
of colour, pH, turbidity (suspended solids) of water),and hence increase the treatment costs. Increased costsare associated with keeping the water intake pointsfree of water hyacinth. For example, Kisumu Munici-pality employs 12 casuals per day, 6 drivers and 6 boatoperators, while Homa Bay municipality engages 2divers at a cost of 1000 Kenya shillings (Ksh) per day.In Homa Bay municipality, water hyacinth builds up 3to 4 times in a week and it takes 3–4 hours to remove it.
Table 3 indicates that water supply capacity inHoma Bay municipality has increased over the studyperiod from 1300 m3 (1986) to 1400 m3 (1998), whilethe price of water has risen greatly from Ksh2 (1986)to Ksh120 (1998) per m3. There were no data for thesupply of water for Kisumu municipality.
The villages bordering the lake have no access to thelake to draw water at times when the beach is heavilyinfested with water hyacinth. Even if they get access tothe water, it is dirty and often smelly because of therotting mats of the weed. This is true for all the 143gazetted beaches along the Kenyan side of Lake Vic-toria which may be infested. The same has been notedin areas around Mwanza and Musoma in Tanzania.
Table 2. Problems associated with water hyacinth in the Lake Victoria Basin
Category Nature of problems Number of respondents
% Respondents
Social Lack of clean water (debris-free)Less access to water points (domestic and livestock use)Societal conflictIncrease in incidence of snake biteDisappearance of the aesthetic value of water bodiesIncrease in disease outbreaks (schistosomiasis, cholera etc.)Reduction of riparian-based tradeMigration of communitiesPercentage of overall responses
13113523
19
5.315.8
5.35.3
15.826.310.515.834.5
Economic Reduced fish catchesIncrease in transportation costsDifficulties in electricity generationDifficulties in water extraction and purificationInterference with irrigation (blockage of canals)Effects on tourismEffects of control on government budgetPercentage of overall responses
8545121
25
32.020.012.020.0
4.08.04.0
45.5
Environmental Decline in diversity and abundance of aquatic life Decline in water qualityIncreased water lossIncreased siltationIncreased potential for floodingPercentage of overall response
42221
11
36.418.218.218.2
9.120.0
136
Health
The hyacinth mats provide breeding grounds formosquitos and other vectors of disease. However,data from Kisumu District (Table 4) indicate declinesin the incidences of malaria (35%) and typhoid (64%)over the study period. The trend in Homa Bay Districtcould not be determined because of sub-division ofthe old South Nyanza District into five districts since1991. There were no comparative data given forMigori District. In both Kisumu and Homa Bay Dis-tricts, there were no reported cases of snakebites. Sta-tistics for Homa Bay for the period 1986–1991 and1992–1998 indicated increased incidence of malaria,whereas there was a decrease in malaria cases inKisumu over the same period. There were more casesof cholera but numbers of typhoid cases fell over thesame periods; and there a slight decrease in amoebi-osis. The statistics on health should be interpretedwith great care as these apply only to incidencesreported to the hospitals or health centres in the dis-trict. There may have been many more cases ofdisease that were not reported to hospitals.
Other socioeconomic impacts
Other socioeconomic impacts reported were loss ofearning opportunities when fishermen could not
access fishing and fish landing sites, as well as inter-ference with fishing gear and clogging of pumps. Rec-reational activities were also affected. Frequent motorbreakdowns and long distance travel in search of unin-fested areas added to costs, likely spoilage or extraexpenditures on preservation. In Uganda, interferencewith electricity generation was reported at the OwenFalls. Extra generation costs, and reduced efficiencywere apparent until the water hyacinth was success-fully controlled.
Environmental impactsThe impacts on the environment were not apparent
and thus not well perceived by most of those inter-viewed among the communities. However, those thataffected the communities directly and posed healthrisks were water quality (foul smell, debris) degrada-tion, increased siltation and potential for flooding.However, other less obvious impacts included inter-ference with diversity, distribution and abundance oflife in aquatic environments. The death and decay ofwater hyacinth vegetation in large masses may createanaerobic conditions and production of lethal gases.
Infestations of water hyacinth affect biodiversity.Dense mats of the weed covering the water surfacelead to deoxygenation of the water, thus affecting allaquatic organisms. It is known that a dense cover ofwater hyacinth enhances evapotranspiration.
Death of water hyacinth mats may influencechanges in the composition, distribution and diversityof aquatic organisms as follows. • Displacement of hydrophytes and depressed algal
biomass (Twongo et al. 1995; Twongo and Balirwa1995).
• Increase in diversity and abundance of some taxa ofmacrofauna, especially at the borders of the weedmats (Wanda 1997).
• Increase in the distribution and abundance ofschistosome (bilharzia) snail vectors e.g.Biomphalaria spp. and Bulinus spp.
• Willoughby et al. (1993) reported that, based onstudies on the Ugandan shoreline of Lake Victoria,mats significantly depressed the diversity of fishspecies and fish biomass. It was subsequentlydemonstrated that fish diversity, particularly smalltaxa, increased along the edge of water hyacinthmats (Twongo and Balirwa 1995). It is evident that quantitative data for many of the
perceived impacts are not available and that if any areavailable they are difficult to analyse, because of themany interrelated parameters that may influence thesocioeconomic status of the riparian communities.
Table 3. Water supply in Homa Bay municipalitysince 1986
Year Quantity(m3)
Price(Ksh/m3)
No. of consumers(water meters)
1986 1300 2.00 500
1987 1300 2.00 520
1988 1300 2.00 550
1989 1300 18.00 600
1990 1300 18.00 650
1991 1300 18.00 700
1992 1400 30.00 770
1993 1400 30.00 800
1994 1400 60.00 950
1995 1400 60.00 1000
1996 1400 90.00 1200
1997 1400 120.00 1500
1998 1400 120.00 1700
Source: Water Engineer, Homa Bay DistrictNote: 78Ksh ≈1 US$
137
It may also be argued that, although water hyacinthposes a serious economic cost to the riparian states, thesame can be said to have presented an opportunity forincome earning to many labourers and manualworkers in the region.
Control Strategies and Status of Control
Although water hyacinth has posed serious economic,social and environmental consequences, there isreason to hope that the control strategies adopted willeventually permit effective management of the weed.
Biological control
Since December 1996, KARI has been introducingNeochetina weevils from Australia, South Africa andUganda as part of a biological control program forwater hyacinth. Community-based lakeside rearingfacilities have produced over 142,000 mostly adultweevils, which were released into the lake at 30 sites in8 districts bordering Lake Victoria. Visual observa-
tions and pre and post-release sampling protocols havebeen used to monitor and evaluate the establishment,spread and impact of the Neochetina weevils on waterhyacinth. Weevils are now firmly established in allaffected areas and have spread as far as 50 km frompoints of release.
Natural enemies on the weed have been observed tohave a significant impact and localised complete sup-pression of resident water hyacinth mats has beenrecorded at all sites including the Police Pier, Yacht club(Kisumu) and Bukoma Pond (Busia) some 24–36months after release. Ecological succession by otherplant species, including hippograss (Vossia cuspidator),papyrus (Cyperus papyrus) and morning glory (Ipomeaaquatica), is now evident in most parts of the lake.
Importation and mass rearing of additional biolog-ical control agents, the moth Niphograpta albiguttalis(previously called Sameodes albiguttalis) and miteOrthogalumna terebrantis, was attempted, but thesedid not establish. Nevertheless, it is recommended thatthese be released in the Lake to augment control byNeochetina weevils.
Table 4. Statistics on disease infection for the Kisumu and Homa Bay districts since 1986
Type of diseases
Malaria Cholera Typhoid Amoebiasis
Year Kisumu Homa Bay Kisumu Homa Bay Kisumu Homa Bay Kisumu Homa Bay
1986 342633 362448 – 40 – 23 – 58716
1987 328021 325370 – 91 – 19 – 63158
1988 259839 276841 – 0 – – – 46012
1989 227756 377128 – 0 – – – 60914
1990 315570 372683 – 0 759 22 – 54641
1991 321942 469242 – 22 312 0 – 64903
Mean 299320 362952 – 26 536 16 – 58064
1992 252363 72775 5 43 252 71 – 7023
1993 19006 68553 – 19 205 152 – 5562
1994 178574 63229 – 38 191 76 – 6518
1995 148467 56692 – 0 107 88 – 4990
1996 177008 55718 – 0 287 114 – 3341
1997 278526 66861 2766 1087 217 113 – 6597
1998 128334 58694 3376 1392 76 76 – 5253
Mean 193454 63227 2049 368 191 99 – 5612
Source: Medical Officer of Health – Kisumu and Homa Bay District.
138
Recent sampling of weevil populations indicatesdensities of between 5.4 and 6.1 weevils per plant.Other observations include:• disintegration of original water hyacinth mats into
smaller mats;• stress on the remaining water plants to reduce them
to seriously stunted weeds;• a general decline in water hyacinth biomass
incapable of flowering and also incapable ofproducing ramets (daughter plants); and
• rotting water hyacinth biomass floating in smallerislands.
Ecological succession
Ecological succession—the progressive displace-ment of one or more species of plants by otherspecies—has made a major contribution to the controlof stationary mats of water hyacinth along the shoresand banks of rivers. In Lake Victoria, pure mats ofwater hyacinth were invaded initially by aquatic ferns/sedges (Cyperus papyrus and Ipomea aquatica) oftento be followed by hippograss (Vossia cuspidator)which invariably eventually dominated and shaded outthe remaining stressed and dying/rotting water hya-cinth. By April 1999, stunted and disintegrated mats ofwater hyacinth and invading weed succession wereclearly evident.
Although water hyacinth will be a permanentfeature in Lake Victoria, currently hippograss and notwater hyacinth forms the dominant weed. The hip-pograss is expected to die once the nutrients fromdying water hyacinth are depleted.
Myco-herbicide development
Fungal pathogens have been known to attack waterhyacinth and so far 32 isolates of fungi have been iden-tified and are maintained in the laboratory. Initial path-ogenicity tests indicate that some of the isolates (suchas Alternaria eichhorniae and Curvularia lunata) maycause over 50% damage to water hyacinth plants underglasshouse conditions. Fungal pathogens have thushelped to also stress the water hyacinth after initialattacks are made by the weevils. Additional fungal iso-lations are still being carried out to increase the level ofattack by obtaining even more virulent isolates. Path-ogenicity tests under ambient environmental condi-tions will be carried out at Kibos, while additionalisolations, identifications, host-specificity, glasshousepathogenicity tests and product formulation work con-tinues at Muguga.
Physical control
Manual removal
The fisherfolk communities around Lake Victoriahave identified key sites for manual removal. Theseinclude fish-landing beaches, ports and piers, irriga-tion canals and water supply points and sources. Fish-landing beaches in most of the affected districts are theprime targets for manual removal operations. Fromover 100 gazetted beaches, some 25–30% had beenseverely infested by water hyacinth at specific periodsduring the year. In 1997, hand tools were distributed.
Mechanical control
Mechanical control operations have so far consistedsolely of chopping and dumping of the chopped piecesof water hyacinth and other weeds into the lake.Regrowth of the chopped weed is likely to take place,especially if most of the natural enemies are destroyedduring chopping. In addition, shallow areas of the lakeare likely to fill up with vegetation, especially alongthe shoreline, leading to drying up and subsequentreduction in the size of the lake. The use of machinesto destroy or remove water hyacinth has limitations,including their inability to move around a large lake.The future of mechanical control options should bereassessed.
Conclusion
The survey and consultations on the impacts of waterhyacinth in the Lake Victoria Basin indicate that, cur-rently, water hyacinth biomass is declining, butslowly. Despite this trend in Lake Victoria, indicationsare that other freshwater bodies continue to beinfested. The precise areas of coverage by the weedcannot be accurately known because of a lack ofappropriate tools and systems for monitoring and ofproper coordination of activities throughout theregion. The range of social, economic and environ-mental problems caused by the weed are generallywell perceived by communities living in the Lake Vic-toria Basin. However, quantitative data are not usuallyavailable to give an idea of the real implications andimpacts on the communities. This presents a challengeto socioeconomists to quantify impacts and thus definethe real problems caused by water hyacinth. A com-prehensive strategy needs to be crafted to address thesocioeconomic and environmental concerns associ-ated with water hyacinth. Further, a coordinatingmechanism needs to be put in place to ensure that acommon approach is adopted in the East Africa and
139
Great Lakes regions to manage water hyacinth andother invasive weeds. For the future, it is recom-mended that a common approach needs to be imple-mented to the management of water hyacinth in orderto enhance understanding and alleviate some of theimpacts associated with infestation by this weed.
References
Bugaari, H., McNabb, T.J. and Moorhouse, T. 1998. Thewater hyacinth problem in Lake Victoria, East Africa.Kampala, Uganda, Aquatics Unlimited, Report WP-010298, 9p.
Freilink, A.B. 1991. Water hyacinth in Lake Victoria: areport on an aerial survey on 12–13 June 1990. In:Thompson, K., ed., The water hyacinth in Uganda:ecology, distribution, problems and strategy for control.Proceedings of a workshop held 22–23 October 1991.Rome, FAO TCP/UGA/9153/A.
Goodland, R. 1995. Uganda: Owen Falls Hydro-project. anenvironmental reconnaissance of the water hyacinthproblem. Report Number S-5043 on a Consultancy Trip,4–6 July 1995. Washington, DC, USA, EnvironmentDepartment of the World Bank.
Labrada, R. 1996. Status of water hyacinth in developingcountries. In: Charudattan, R., Labrada, R., Centre, T.D.and Kelly-Begazo, C., ed., Strategies for water hyacinthcontrol. Report of a Panel of Experts Meeting, 11–14September 1995 Fort Lauderdale, Florida, 3–11.
Lowe-McConnel, R.H., Crul, R.C.M. and Roest, F.C. 1992.Symposium on resource use and conservation of the GreatLakes held in Bujumbura, Burundi, 1989. Stuttgart, E.Schweigerbatsche Publ. for The InternationalAssociation of Theoretic and Applied Limnology 23,128.
Mailu, A.M. 1998a. Water Hyacinth Information BulletinNo. 1, June 1998.
— 1998b. Water Hyacinth Information Bulletin No. 2,December 1998.
Taylor, A.R.D. 1993. Floating water weeds in East Africa,with a case study in northern Lake Victoria. In:Greathead, A. and de Groot, P., ed., Control of Africa’swater weeds: Series Number CSC (93) AGR-18.Proceedings of a workshop held in Harare, Zimbabwe,June, 1991, 111–119.
Twongo, T. 1993. Status of water hyacinth in Uganda. In:Greathead, A. and de Groot, P., ed., Control of Africa’swater weeds: Series Number CSC (93) AGR-18.Proceedings of a workshop held in Harare, Zimbabwe,June 1991, 55–57.
— 1996. Growing impact of water hyacinth on near shoreenvironments of Lakes Victoria and Kyoga (East Africa).In: Johnson, T.C. and Odada, E., ed., Proceedings of asymposium: The limnology, climatology and paleoclima-tology of East African Lakes, 18–22 February 1993, Jinja,Uganda.
Twongo, T. and Balirwa, J. 1995. The water hyacinthproblem and the biological control option in the highlandlake region of the upper Nile Basin — Uganda’sexperience. Unpublished paper presented at The Nile2002 Conference. Comprehensive Water ResourcesDevelopment of the Nile Basin, held 13–17 February1995 Arusha, Tanzania.
Twongo, T., Bugenyi, F.W.B. and Wanda, F. 1995. Thepotential for further proliferation of water hyacinth inLakes Victoria, Kyoga and Kwania and some aspects forresearch. African Journal of Tropical Hydrobiology, 6, 1–10.
Wanda, F.M. 1997. The impact of water hyacinthEichhornia crassipes (Mart.) Solms (Pontederiaceae) onthe abundance and diversity of aquatic macro-invertebrates in northern Lake Victoria, Uganda. MScThesis, Delft, The Netherlands, International Institute ofInfrastructural, Hydraulic and EnvironmentalEngineering.
Willoughby, N.G., Watson, I.G., Lauer, S. and Grant, I.F.1993. An investigation into the effects of water hyacinthon the biodiversity and abundance of fish andinvertebrates in Lake Victoria, Uganda. NRI ProjectNumber 10066 A0328. Visit Number 2007(S).
140
Biological Control of Water Hyacinth by Neochetina eichhorniae and N. bruchi
in Wenzhou, China
Lu Xujian*, Fang Yongjun†, Song Darong† and Xia Wanqing‡
Wenhzou is located in the southeastern Zhejiang Province. Water hyacinth was introduced into Wenzhou in the1960s. Now the weed has become a bio-disaster. Chemical and manual control of water hyacinth cannot controlthe weed. In co-operation with the Biological Control Institute, Chinese Academy of Agricultural Sciences,Neochetina eichhorniae was introduced and released at four sites (Wutian, Liushi, Yaoxi and Kunyang) in1996. The area of each site is about 1000 m2 and 1000 adult weevils were released at each site. The weevilshave established and spread their population well at Wutian and Liushi. However, although they becameestablished one year after release, there were no weevils at the other two sites, because of the effect of someunknown factor. The study in Wenzou showed that the maximum total numbers of weevils occurred in mid-June each year. Adult numbers were highest in November, larvae in mid-June and pupae in early July. From1996 to 1999, the area and height of water hyacinth had been reduced greatly at Wutian and Liushi.
* Wenzhou Academy of Agricultural Sciences, XinqiaoStreet, Wenzhou, Zhejiang, 325006, China
† Wenhzou Agricultural Bureau, Shouyi Qiao, JiushanbeiluRd., Wenzhou, Zhejiang, 325006, China
‡ Ouhai District Agricultural Bureau, Shouyi Qiao,Jiushanbeilu Road., Wenzhou, Zhejiang, 325006, China
142
Summary of sessions 1 and 2
Session 1, which was of keynote presentations, was chaired by Martin Hill, and session 2, containing general papers on the biological control of water hyacinth, by Mic Julien. This report was prepared by Ted Center.
Sessions on Tuesday 10 October opened with three excellent keynote presentations,the first by Mr Mic Julien on ‘Biological control of water hyacinth by using insects in theworld’, the second by Dr R. Charudattan on ‘Biological control of water hyacinth byusing pathogens in the world’ and the third by Dr Ding Jianqing on ‘Water hyacinth inChina: its distribution, problems and control status’.
Mic Julien presented a review of water hyacinth biology and its status as a worldwideproblem, using examples from Papua New Guinea, Thailand and Zimbabwe. Thisincluded a historical perspective on the introduction and spread of water hyacintharound the world beginning in the early 1900s and culminating in the recent widespreadinvasions in Africa. Mic then followed by defining ‘classical’ biological control and heoutlined the steps in a classical biological control project. He then listed new proposedagents (as outlined by Cordo in the previous workshop) and reviewed the biology andimpact of many of the agents in current use.
Perhaps most importantly, Mic explored the factors that relate to successful control,factors that accelerate success, and factors that limit control. Factors that he associatedwith successful control included: presence in tropical and subtropical areas; infestationsmanifested as monocultures in free-floating mats (which are able to sink when dam-aged); and mats that are stable (i.e. undisturbed) over long periods. Factors proposed thatmight accelerate control included wave action, reduced growth (by the actions of bio-control agents, which allows plants to be flushed out of the systems), and high nutrientstatus (as it relates to the production of high quality plants thus enhancing insect popu-lation growth). Factors that limit control included removal of mats by herbicidal ormechanical means (thus disrupting agent populations), shallow water (damaged plantsunable to sink), ephemeral water bodies, toxicity effects in polluted waters, low temper-atures at high-altitude, temperate sites, and high nutrients at temperate sites. Theseobservations represent important hypotheses needing further testing, but this delineationrepresents a first attempt to come to understand the variable nature of successful biocon-trol.
Mic concluded by providing recommendations for improving water hyacinth biolog-ical control. He suggested that existing agents needed wider use, that new agents had tobe identified, that better resources were needed to support projects, that stakeholdersneeded to be educated about biocontrol, that more collaboration with weed scientists andmanagers was needed, and that catchment-specific integrated weed management plansneeded development.
Dr Charudattan began his presentation by stating that, in his view, pathogens would bemost useful as biopesticides as opposed to ‘classical’ biological control agents. He
143
emphasised that pathogens should be explored fully and fairly so as to assess their poten-tial as biopesticides in integrated weed management systems (as opposed to ‘standalone’ agents). He felt that the use of ‘local’ pathogens avoided the quarantine problemsassociated with exotic pathogens. He noted that no new pathogens had been identifiedin the past 20 years, despite numerous surveys. The only viable ‘classical’ candidate thathe identified was the rust fungus Uredo eichhorniae. He felt that the bioherbicidalapproach could produce quick and acceptable levels of control, a need generally per-ceived by aquatic weed managers and a prerequisite to widespread adoption.
Charu then enumerated and prioritised known water hyacinth pathogens. He consid-ered Cercospora piaropi to be one of the most promising, being widespread with manyvirulent strains. Acremonium zonatum also showed some promised, as did Alternariaeichhorniae and Myrothecium roridum (although the availability of strains with differentlevels of virulence of the latter two species is not yet clear). Rhizoctonia solani, whichis highly virulent, had not been considered previously, because of its lack of host spe-cificity but Charu noted, in the present regulatory environment, its use is now possible.Many other species of less widely developed pathogens were noted.
Charu summarised by suggesting ways to make pathogens more effective by over-coming environmental constraints on their growth and efficacy. Once developed as bio-herbicides, they can be: applied with low rates of herbicide or with adjuvants; multipleapplications can used to increase inoculum loads; applications can be timed to maximiseimpact with insect agents; novel formulations may be used to provide humidity on theleaf surface, to protect the inoculum from solar radiation, or to promote hyphal penetra-tion of the leaf; or they can be combined with other pathogens or several strains of thesame pathogen to provide consistent performance. The overall objectives in theseapproaches should be aimed at developing a bioherbicide that can be used with existingagents and improve the overall effectiveness under different control scenarios so as to‘knock back’ the weed as opposed to ‘knock down’. One recommendation was to applyCercospora piaropi in combination with Myrothecium roridum using a surfactant.
In conclusion, Charu felt that it was indeed a challenge to develop an effective andpractical bioherbicide, but that the challenge could be met given the bio-friendly postureof modern regulatory agencies. However, newer, innovative approaches were requiredto meet particular needs.
Ding Jianqing provided a history of water hyacinth in China and noted that it is widelyutilised as animal food so not everyone regards it as a problem. Water hyacinth wasintroduced in Taiwan in 1903 from Southeast Asia and then to the mainland in 1930. Itwas introduced into almost every province in the 1950s and 1960s for use as animalfood. It became more damaging as nutrient pollution increased during the 1990s. Waterhyacinth is now distributed in 17 provinces and is regarded as a problem in 10 of them,mostly in the south. China spends 100 million RMB annually to manually control waterhyacinth. The problems caused by water hyacinth are generally the same as in othercountries (however, one that may be unique to China was that it provided hiding placesfor criminals – nicely illustrated with a self portrait surrounded by the weed!). Ding pre-sented a very interesting example of the impact of water hyacinth on the biodiversity ofDianchi Lake. The number of plant species known in the lake declined from 16 beforethe introduction of water hyacinth to 3 afterwards. Although invasive species are oftenblamed for adversely affecting biodiversity, this is rarely documented.
144
Ding emphasised that utilisation does not solve the problem and should not be con-fused with control. Mic Julien had earlier emphatically stated this as well.
Biological control was initiated in China in 1995 and, after demonstrating the effectsof the weevils in small-scale tests, Neochetina bruchi and N. eichhorniae were releasedin 1996. Significant control is now being realised in Zhejiang and Fujian provinces.More recently, the mirid Eccritotarsus catarinensis has been released and surveys forpathogens have begun. Chinese researchers have also demonstrated that integrated con-trol, using herbicides with the weevils, provides better control than either alone. In thefuture they plan to introduce the weevils into new areas, to introduce new insects, and todevelop bioherbicides.
One of the most profound statements of the meeting was made by Ding in his con-cluding remarks. He stated that control of water hyacinth is not about technology, butrather about politics.
In the general session that followed the keynote presentations, Dr Martin Hill pre-sented the first paper, an introspective look at biological control of water hyacinth inSouth Africa focusing on constraining factors, success, and new courses of action. SouthAfrica has introduced more agents (6 total) than any other country, yet water hyacinthis still a problem. He identified several of the same factors enumerated by Julien as pos-sibly being inhibitory, plus some others: diverse climate (ranging form high altitude tocoastal Mediterranean and coastal subtropical), eutrophication (at some sites 100% ofinflow is treated sewage effluent), herbicide interference (direct toxic effects as well asremoval of bioagent habitat), hydrological parameters (small, shallow systems), andlimited releases (small, inoculative releases as opposed to mass rearing and release).
Martin then presented a case study from an oligotrophic system at New Year’s Damwhere biological control agents reduced water hyacinth coverage from 80% in 1990 toless than 10% in 1994. Thus, even though biological control is not universally effectiveeverywhere, it is effective in the proper circumstances. He concluded with the feelingthat water body managers expected too much from biological control, that the problemof eutrophication needed to be addressed, and that better integration with other controlmeasures was needed. He also echoed the sentiments of almost every participant thatadditional biological control agents were needed.
Tom Moorehouse then provided an overview of the biological control activities in theKagera River Headwaters in Rwanda. The Kagera River feeds into Lake Victoria on thewestern side and is the source of much of the water hyacinth entering the lake. In col-laboration with Ugandan researchers, Clean Lakes, Inc. is setting up weevil rearingsystems and training the Rwandans in the use and application of biological control.
Dr Yahia Fayad presented great news about the biological control of water hyacinthin Egypt. Although the weevils were studied in the early 1980s, the Egyptian Govern-ment refused to grant permission for their release. However, the previous working groupmeeting provided the stimulus for them to finally issue approval. As a result, bothspecies were released during August 2000 when 2000 weevils were liberated at MarioutLake and 4000 at Edko Lake.
The afternoon session concluded with another presentation by Dr Hill on the biolog-ical control of water hyacinth in Malawi. Water hyacinth was found at Lake Malawiduring the late 1960s, and by 2000 it was scattered throughout the country. Rearingsystems for Neochetina weevils have been set up at two facilities along the Shire River.About 200,000 weevils have now been reared and released, 47% in the upper Shire
145
River, along with the water hyacinth mirid Eccritotarsus catarinensis and the mothNiphograpta albiguttalis. The mite Orthogalumna terebrantis was already present. (Incontrast to Mic Julien’s conclusion that these mites are ineffective, they seemed to bequite damaging in this area.) The weevils have established and monitoring is continuing.Some reductions in water hyacinth have already been noted but new infestations areappearing elsewhere. Nonetheless, Martin was optimistic.
146
Summary of session 3
Session 3, ‘Biological Control – Pathogens’, was chaired by Garry Hill. This report was prepared by R. Charudattan.
Three papers were presented in this session. The first, entitled ‘An International Col-laborative Program to Investigate the Development of a Mycoherbicide for Use AgainstWater Hyacinth in Africa’, was presented by Roy Bateman (CABI Bioscience, UK), onbehalf of the International Mycoherbicide Programme for Eichhornia crassipes Controlin Africa (IMPECCA) team. He gave an overview of the IMPECCA program, in whichthe following groups are participating: CABI Bioscience, UK; the International Centrefor Research on Agro-forestry, Kenya; the International Institute for Tropical Agricul-ture, Benin; the Danish Institute of Agricultural Sciences, the Plant Protection ResearchInstitute, South Africa; and the University of Mansoura, Egypt.
The project will build upon existing studies of mycoherbicides that have been carriedout in Egypt and Zimbabwe. Explorations will be undertaken to find pathogens native tothe African continent and the pathogens will be characterised by cultural and morpho-logical studies. Weed biotypes will be characterised by a molecular (AFLP, amplifiedfragment-length polymorphism) technique. Preliminary results indicate that the waterhyacinth biotypes examined to date are similar—a promising sign that host genotypicdifferences will not be a complicating factor in this mycoherbicide program. Formula-tion studies are under way with emphasis on an oil-based formulation that could besprayed like a conventional herbicide. An application technology based on the suc-cessful locust-control program (LUBILOSA program) will be evaluated for implemen-tation of a mycoherbicide for water hyacinth. Pathogenicity tests are also being carriedout with the following fungi found in the participating countries: Alternaria alternata, A.eichhorniae, Acremonium zonatum, Cercospora piaropi, Rhizoctonia solani, andMyrothecium roridum.
The IMPECCA program will be promoted and widely publicised. The Water HyacinthNewsletter published by CABI Bioscience will be continued under this program. A tech-nical bulletin, edited by Roy Bateman, will be made available to interested individualsand agencies. A web site (http://www.impecca.net) has been set up and is running.Another key output will be the strengthening of technical capability and linkages withinAfrican national programs to undertake biological control of weeds.
Harry Evans (CABI Bioscience, UK) then presented a paper (by Evans and Robert H.Reeder) on the fungi associated with water hyacinth in the upper Amazon basin, and theprospects for their use in biological control. He pointed out that few fungi have beenreported on water hyacinth in South America, most of the fungal taxa reported comingfrom the Palaeotropics rather than the Neotropics. He detailed his observations fromsurveys he undertook in 1998 and 1999 in the upper Amazon basin of Ecuador and Peru.The objective of the surveys was to document both the distribution of water hyacinth andits associated microbial pathogens. Ecuador and Peru are within the purported center of
147
origin or diversity of water hyacinth and no surveys had previously been made in thisarea.
Three groups of fungi were collected: biotrophs, colonising green tissues without sig-nificant external symptoms; necrotrophs, causing prominent leaf lesions; and thoseassociated with tissues damaged by insects (Taosa spp., Thrypticus spp.), typically onthe petioles. Several fungi that belong to known plant pathogenic genera were isolatedor recorded on the collected specimens. These remain to be tested. These surveys haveyielded several interesting fungi, including some that appear to be new species. In myview, this collection represents an important resource and it would be very worthwhileto test and evaluate these fungi.
Yasser Shabana (Mansoura University, Egypt; currently and temporarily at Universityof Hohenheim, Stuttgart, Germany) reviewed his work on Alternaria eichhorniae con-ducted during the past 16 years. This work (by Shabana, Elwakil and Charudattan) hascovered a large area of the subject, including the initial surveys in the Nile Delta (Egypt),host-range and safety tests of A. eichhorniae, investigations on several aspects forimproving the efficacy of this mycoherbicidal agent i.e. sporulation, phytotoxin produc-tion, bioherbicide formulations, determination of optimum epidemiological conditionsfor disease incidence and disease severity, and physiological and ultrastructural studies.This work has clearly established this fungus as a leading candidate for mycoherbicidedevelopment in the African continent. Shabana presented results from his most recentwork on formulation and field-testing of this fungus, concluding that, for best results,spore or mycelial inoculum of this fungus should be formulated in an invert oil emulsion,along with phytotoxic metabolites produced in culture. The phytotoxic fractionspromote disease development, while the invert emulsion protects the fungal inoculumfrom dehydration.
Overall, this was an interesting and informative session highlighting the importance ofpathogens as biological control agents of water hyacinth.
148
Summary of session 4
Session 4, ‘General’, was chaired by Ding Jianqing. This report was prepared by Harry Evans.
Two talks were presented in this session, both covering information technologyrelating to water hyacinth. Firstly, Garry Hill introduced the idea of a multifunctionalWater Hyacinth Resource Manual which, as well as being an information source, couldalso be used both as a training tool and as a guide to decision-makers. The manual wouldbe focused specifically on developing countries in Africa and would be prepared in theform of a comprehensive, practical and authoritative directory, by an editorial teamunder a professional editor. It was envisaged that there would be a workshop aroundmid-2002 to finalise the contents of the document.
The session participants considered this to be an excellent initiative and a potentiallyvaluable and much-needed resource, although the ambitious subject area could mean thatthe manual may run into several volumes. It was further suggested that the documentcould be in a loose-leaf format and that the initial drafts should be presented on a web site.
A lively discussion ensued over the costs involved and the potential donor sources. Luis Navarro then presented a proposal to aid decision-making through the establish-
ment of a Water Hyacinth Information Partnership (WHIP). While WHIP would beaimed at presenting information to African and Middle Eastern countries, its principalmission would be to link this information rather than to create it, facilitated by an Infor-mation Exchange and Networking Mechanism (IENM).
Thus, the partnership could be employed for the early detection of water hyacinthinfestations and thereby stimulate decision-makers to respond more rapidly to theircontrol or eradication. In the past, such decisions have not been taken until weed levelshave become critical and this has been the experience in 21 countries since the 1980s.Thus, there are political reasons for WHIP: to raise donor awareness of the problem andto serve as an information source to decision-makers; as well as technical reasons—toprovide data on the spread, socioeconomic costs and integrated control strategies.
It was proposed that this concept could be extended to other invasive weeds and thata conference should be organised to present the initiative, particularly to potentialdonors. The general view of the session participants was that the projected budget(US$1.5 million) was excessive for the envisaged outputs.
149
Summary of session 5
This session, ‘Biological Control – Insects’, was chaired by Dr Lu Qingguang. This report was prepared by Mic Julien.
Ted Center presented thought-provoking ideas and early studies that assess thechanges in competitiveness of water hyacinth when attacked by biological controlagents. This is work prepared by Center, Van and Hill. He suggested that this techniquecould be used to select the most damaging agents for release. The idea has considerablemerit, especially if it improves our predictive capacity and thus saves time andresources. The experimental design may need to pay attention to nutrient levels.Changes in nutrition may alter agent impacts or the plant’s ability to withstand or com-pensate for damage. The experimental design is already large, and adding nutrients as atreatment will increase its size considerably.
The Oberholzer and Hill paper, presented by Martin Hill, reported the results ofstudies on Cornops aquaticum, a potential control agent that has long been waiting in thebackground. It appears that this insect is specific to Pontederiaceae. That C. aquaticumtook a bite of banana during testing is undoubtedly an aberration. Because of the impor-tance of bananas, the South African Plant Protection Research Institute will study thisphenomenon further. Even if later studies show that banana is not a suitable host, theoriginal results should be reported and decision-makers may misconstrue this. In thisrespect we are all obliged to educate the people who sit on committees and make deci-sions about what is and is not safe to import and release.
Andrew Mailu described the impacts of water hyacinth on life on, in and around LakeVictoria. He indicated the importance of prior experience in decision-making. Kenyamoved quickly to embrace biological control of water hyacinth following experiencewith salvinia. He indicated that the project on water hyacinth was bigger than just waterhyacinth and embraced the whole of catchment management, and that the success of bio-logical control of water hyacinth will have far-reaching effects in the area. He discussedthe importance of obtaining reliable scientific and socioeconomic data to describe thecosts of problems and the benefits of solutions.
John Wilson, PhD student, outlined progress with modelling the water hyacinth andthe impact of controls. His study attempts to bring together data and ideas worldwide,and to investigate some areas of plant–insect interactions. The broad aims are to identifyareas needing research, identify the factors that limit control and to help make area-spe-cific predictions about control.
Martin Hill gave an outline of activities in Benin and Zambia, and indicated that a bio-logical control project was starting in Burkina Faso. Excellent control has been achievedin several locations in Benin, with the weed infestations less than 5% of their formersizes. The project included community involvement with very beneficial results. Theproblems in the Kafue River (95 km infested) in Zambia remain. The weevils werereleased in the 1990s and are established. Niphograpta and Eccritotarsus were recently
150
released. Nutrient levels are being sampled. Insufficient work is being conducted toassess progress. However, the continued presence of the weed suggests that the weevilsand the mite are insufficient. This infestation presents itself as an ideal location to under-take a large-scale integrated management trial.
Godfrey Chikwenhere outlined the status of biological control of water hyacinth inZimbabwe and the continued involvement of politics in the decision-making processes.Regardless, and this has been shown in other situations, he indicated that once theweevils were widely established they were able to contribute to control despite otherantagonistic controls that were imposed. He also indicated the importance of consid-ering other aquatic weeds at the same time, so that weeds such as water lettuce did notimmediately take up space made available after water hyacinth was controlled.
Finally, Eric Gutiérrez outlined biomass and productivity studies he had conducted inMexico with Gomez and Franco, preparatory to determining appropriate controlmethods for each site. Such studies have rarely been carried out before purchasingmechanical removal machines and hence the tropics are littered with expensive andlargely useless machinery. He outlined successful mechanical removal operations con-ducted at a number of impoundments and described the current ongoing efforts ofreleasing 4000 adult weevils per month into other locations. It would be instructive toknow the costs of the mechanical removal procedures.
This session of seven presentations was varied and very stimulating. We enjoyed talksincluding assessment of potential agents, development of predictive capacity, reports ofsuccessful control (biological and mechanical), reports of apparent failure of agents tocontrol the weed, the need for total catchment management and to collect sound data thatdescribe the costs of weeds and benefits resulting from control. I thank the presentersand the audience for the lively discussions.
151
Summary of session 6
Session 6, ‘Integrated management’, was chaired by Roy Bateman. Martin Hill prepared this summary.
There were two talks and one video in this last session of the workshop. In the firstpaper, Gasper Mallya presented an update of the water hyacinth biological controlprogram in Tanzania. As with elsewhere on Lake Victoria, fantastic results have beenachieved with the release of Neochetina eichhorniae and N. bruchi. Much of the successhas been ascribed to having dedicated staff and enlisting participation of the local fishingcommunities. This has resulted in the release of large numbers of the weevils and hasensured that collection and redistribution of the weevils to other localities has occurred.Currently, as a result of biological control, water hyacinth is not considered a problem onthe Tanzanian shores of Lake Victoria.
Several initiatives have been put in place in Tanzania to ensure sustainable control ofwater hyacinth. Firstly, utilisation of the weed has been prohibited, thus preventing thepotential spread of water hyacinth to other freshwater bodies in Tanzania. Secondly, fish-ermen are required to report basic information on water hyacinth infestations on the laketo the relevant authorities. This includes general observations on the size of the plants, butmore importantly the occurrence of any new infestations to aid in the early detection of thespread of the weed to new sites. Thirdly, releases of the weevils in the Kagera River, whereit runs through Tanzania, will facilitate the biological control of water hyacinth on LakeVictoria.
The water hyacinth program in Tanzania, as with the programs in Kenya and Uganda,has been extremely successful in reducing the environmental and socioeconomic threatposed by water hyacinth on Lake Victoria.
In the second presentation, Roy Jones showed a video and then presented a paper on theintegrated control of water hyacinth on the Nseleni River and Lake Nsezi in northern KwaZulu–Natal, South Africa. This program was inexpensive and highly successful and hasresulted in the return of a number of endangered waterfowl species to this conservationarea. Roy stressed that having realistic expectations of what can be achieved was the firststep. The second step was to obtain commitment from the communities along the system.Without this commitment, the program was likely to fail. The third step was to divide thesystem into manageable units from the top of the system, downstream. However, follow-up in units after initial clearing of the weed, to control regrowth, was vital. This single mostimportant factor in the success of this project appeared to be having a dedicated managerwho spent considerable time on the system and who could coordinate the control efforts.
This Nseleni River program was well planned from the start. It relies on biologicalcontrol but uses herbicidal control and mechanical intervention (the use of cables acrossthe river to prevent the movement of water hyacinth into previously cleared areas) in acoordinated manner. In addition, nutrient control through the upgrading of a wastewatertreatment plant in the upper catchment has facilitated the control efforts. Furthermore, the
152
public awareness campaign through the production of a video, the involvement of the localcommunity around the system and the fact that the program has been well documented hasensured that this is possibly the best example of integrated management of water hyacinththat we have.
